What's the Ideal Team Size?

Read Matt’s history of Automattic: “On Building Automattic.”

On today’s episode, Automattic editor Mark Armstrong interviews our usual host Matt Mullenweg to discuss the history of the company and how its distributed culture emerged from conditions that many startups face. They go deep on the tools and processes Automattic has developed to keep everyone connected, even if they’re scattered across the globe.

The full episode transcript is below.

***

Mark Armstrong: Hi everybody. Thanks for joining the Distributed podcast. I’m not Matt Mullenweg, I’m Mark Armstrong. I’m the founder of Longreads, which is part of Automattic, and I’m on the editorial team working with Matt on the Distributed Podcast.

So today I wanted to take a step back from the interviews Matt’s been doing and find some context for how Matt got here in the first place, how he became interested in distributed work, and it all starts with the history of Automattic. So that’s what we’re going to talk about today. We’re going to talk to Matt about how he got here, how he actually decided to build a company that had no offices, and what worked and what didn’t. Thanks for listening.

Now where are you right now?

Matt Mullenweg: I’m actually in San Francisco, California.

Mark: And I am in Seattle. This is basically how we record this podcast for people who are remote. We have a Zoom connection, we use GarageBand, a tool called Zencastr, and we put it all together and it sounds like we are having this intimate conversation right next to each other. But this is a very good symbol of how distributed work has changed and the technology has changed.

Matt: Yes. 

Mark: So take me through the very beginnings of Automattic and how you ended up with a distributed work model in the first place.

Matt: Well at the very beginning I had moved from Houston to San Francisco, actually, to take a job with CNET Networks. And they were actually an early adopter of WordPress. So they offered me a job as a product manager to drop out of college in Houston. I drove across the country with my mom and got a little apartment in San Francisco.

Mark:  Did you apply for the job?

Matt: So what had happened was, I had said if I hit number one on Google for the search term “Matt” I would shut down my site and go out on top, like MJ. So that happened. Partially because most WordPresses included a link to the main developers, of which I was one. And they all said “Matt” and so those links started helping me rank first for Matt. So I just replaced my site with a little black screen that said this happened and I came out to San Francisco. Because this got some news when this happened. A bunch of people reached out.

Mark:  So at this point, you had already started WordPress. Did you already have in the back of your head the idea of building a company around WordPress or were you thinking, well this CNET thing will work out for a couple of years and see how it goes?

Matt: I’m guessing this was probably 2004. At the time, there was no indication that anything related to WordPress could ever make money. [laughs] It was very much a voluntary project. And there weren’t really examples, certainly not as there are now, of consumer open source becoming a commercial thing. But what I did know was I was already collaborating with folks all around the world having a blast and using basically all my free time to get online, hang out in IRC, which was kind of like an early pre-Slack chat system, and basically code with these people who were also passionate about making publishing easier.

Mark: So you’re doing all this work on something that you love and then you’re commuting into an office in San Francisco. What was the office experience like? Was this the moment where you were like, “I never want to work in an office again so I will never do that?”

Matt: No, the opposite. It was amazing. [laughs] CNET was one of the first companies to actually set up a headquarters in San Francisco. Most of the major companies were and are down in Mountain View or Cupertino or some place else in the Bay Area… San Jose… But I looked very, very close to the CNET office and then when I moved in, my patron there, a guy named Mike Tatum, got me this really cool office right above the door, like a corner office. 

And I remember this VP came by and in the course of the conversation, “Hey, where are you from, how are you doing, what you going to do here,” etcetera, he started to look around and be like, “Man, this is a nice office.” You can guess what happened next. Two weeks later I come to the office and there’s a note on my door. “Oh we’re doing an office move.” And I got moved to an interior office in the same section, still on the third floor, but more of an interior, non-window office. 

But then there was a general re-org in the group I was in, and that whole floor got taken over by GameSpot. That was doing really well. And they literally moved my group into the basement, [laughter] the very, very first area. So now I was not just no-window, I was in this very small area and that was around the time I think I started to think this might not be for me.

Mark: So you’re putting together the idea for Automattic as a for-profit company that then would contribute to the WordPress open source project. So you’re viewing this as an opportunity to merge the two ideas of the non-profit, open-source side and a for-profit company that feeds into it. Is that the original idea?

Matt: So the original idea I actually pitched at CNET. They had a lot of cool domains, famously com.com, download.com. And one day I got a list of all their domains. They had probably a thousand from different acquisitions and stuff. And I went through them and I was like “All right, well, could be cool.” And I came across one called Online.com. 

So the idea in the pitch I made to them was let’s make a version of WordPress that anyone can start with. So instead of having to know PHP and configure a database in FTP and all these things, let’s make it self-serve, where you can just click a few buttons and get one. That could be Mark.online.com. It could be your online home and you could have a blog and a profile and all this cool stuff. So I really pitched that quite vigorously.

But at the time what was going on was this colossal battle between blogs and traditional publications and two, every single internet giant had a blogging system but they were scared to call it blogging. So Yahoo had one called 360, AOL had Journals, Microsoft had something called Live Spaces, and then of course Google had Blogger. And so it appeared that all the internet giants — and those were the internet giants at the time — all had something in this space and CNET, one, thought they couldn’t compete, and two, thought that blogs were just gonna be like noise and politics and mess and junk and they were really going to bet on professional publications. 

They were at the time locked in — News.com, which was a tech publication, was battling — or ZDnet was battling with Gizmodo, which was one of the early blogs started by Nick Denton and the Gawker Network, and Engadget from Jason Calcanis and Pete Rojas and those folks were also just getting started. So they saw themselves locked in this epic battle of professional versus amateur and didn’t want to do with anything related to blogging. 

So I said “Well, I have to do this. I’m going to leave and start this.” And CNET very graciously invested in what became Automattic but asked if I could stay a few months to finish up some of the projects I was working on. So I did. But in those few months, I went ahead and started Automattic. I already knew the people I was working with as the other lead developers of WordPress and started trying to convince them to leave their jobs. [laughs] 

And the first one actually was a fellow named Donncha Ó Caoimh, who was over in Blarney, Ireland. And he had started a different fork of the software that became WordPress. His was called B2++, I believe. And we decided to merge that with WordPress and make that WordPress Multiuser, which later became WordPress Multisite — basically a version of WordPress that was multitenant, that could have lots of people clicking a few buttons and starting a blog. And I want to say the second was either Ryan Boren or Andy Skelton. And I was technically the third because I stayed a few extra months to finish up things at CNET, and then we started Automattic.

Mark: Already you’ve built the beginnings of this company and you’ve already got employees that are spread out all over the world. So was that simply, “Okay well, these are the people I trust and want to work with and so we’re going to build a company that way,” or was there more deliberation around that?

Matt: One aspect was definitely having no money. [laughs] I was just paying Donncha and these other folks out of my CNET salary basically, and for servers and things, starting to rack up some credit card bills. So it was very much like — well, moving is really expensive [laughs] and we’re already working together, and, in fact, those were the days when we’d all work very, very long hours. So if you’re working 10 to 16 hours a day, you’re overlapping quite a bit with someone who is, say, in Ireland. And we had this awesome, almost relay thing where I’d work until night and then he would wake up and start working. And then I’d sleep, wake up, and there’d have been a whole cycle of things done to the software. It actually allowed us to iterate very quickly and brought in some asynchronicity very early on to our interactions.

Mark: At this point, how are you communicating with each other?

Matt: Probably we used AIM. AOL Instant Messenger, which is now shut down, IRC, and we were either still on SourceForge, which was kind of like an early GitLab or maybe at this point had switched to hosting the code on WordPress.org.

Mark: So this is a lot of baseline foundational communication that sounds very familiar to me even in current day Automattic.

Matt: We actually still run IRC servers to this very day. [laughs] And some people like to use them.

Mark: Yes.

Matt: But the evolution at the time, AIM was definitely the most popular chatting platform. At some point we switched to Skype again for chat, not really for voice. But I do remember some of those early voice calls where we, for the first time, would talk to each other and have what maybe [were] our first meetings. And the average meeting cadence at that time was probably three per year. We didn’t do a lot of them. It was really very much a written communication style.

Mark: And you hadn’t met in person yet.

Matt: None of us had met in person. Around that time was when I hosted the very first WordPress meetup in the world, which ended up being eight people at an Indian place, which is still there on Third Street, called Chaat Cafe. Those eight people ended up being pretty interesting. There was Chris Messina, who at the time was involved with Drupal Project, and would later go on to invent the hashtag. Scott Beale of Laughing Squid, which is now a large WordPress webhost. Om Malik, who was a journalist at Business 2.0 and later started GigaOm, became one of the earliest WordPress users. And someone who I met for the first time, which was Ryan Boren, who was one of the first major contributors to the core of WordPress. And he was, I want to say, an embedded systems engineer at Cisco. 

Mark: There is a point though where you are going to then bring Automattic to investors and raise some money. How did they react to this idea that you didn’t have an office, [where] everyone was all over the place?

Matt: We didn’t really plan to raise investment in the beginning, we were just focused on making money to be honest. So we thought that we could have add-ons for WordPress.com. I think some of our early ones were custom CSS for customization and domain names, so this would allow people to be a business model. This idea of ringtones for blogs so you could buy customizations or add ons, actually not dissimilar to what we do today. And then we started to also do partnerships with different hosting companies who provided hosting and then could pay us for essentially new customers.

So that was the early business model and that started to make a bit of money, enough where we could each take a modest salary of a few thousand a month. And that was the plan. We didn’t have a ton in the bank, there was enough for a few of us to work on WordPress full time. And that was really just a dream. I mean, the company was created to have a place where we could be paid to contribute to open source and so we were all happy as clams.

Mark: It seems like you had gotten far enough along with the growing group of contributors and employees that you were already proving that distributed work was working. So by the time you ended up bringing in outside investors or other partners, that wasn’t really a question anymore. Is that right?

Matt: I met some different investors, folks like Tony Conrad and others, but I was at this point 21, [laughs] so, pretty young. And that was the era, pre-Zuckerberg, pre-all these other things, where you brought in adult supervision. So the model was really Eric Schmidt. Adult comes in and professionalizes the business. 

And I didn’t really want that until Om Malik introduced me to a fellow named Tony Schneider, who had sold a company called Oddpost to Yahoo It was the original Ajax Gmail predecessor that allowed you to have a very web-application feel in the browser. They were the first to do that, that I had seen. But he wasn’t going to stay at Yahoo. So Om had done a cover story on the company for Business 2.0, and he said, “Hey you’ve got to meet this kid from Texas.” And he told me, “Hey you’ve got to meet this guy Tony, he’s not going to stay at Yahoo too long.” 

So we met and really, really hit it off. Tony is an amazing individual. He came from Switzerland, moved to California, ended up going to Stanford, had a few start ups under his belt, including this very successful exit at the time to Yahoo, and was a CEO. And I was like, “Ah, this guy is my business soulmate.” If this is the adult, [laughs] I can totally bring in an adult — this guy is amazing. And I learned so much from him.

That was with the expectation, that Tony would join the following year when he was able to leave Yahoo. [He] ended up raising the first round for Automattic, which was about $1.1 million from folks who we actually still work with today, like Phil Black at True Ventures, Doug Mackenzie, Kevin Compton at Radar — so that first early round. And I was the CEO for I forget how long, the first however many months before Tony joined, and then we started working together.

Mark: Was there ever any pressure to eventually — “Okay, this is great, I love distributed work but we also need some physical offices?” How much pressure had come in, in those early days?

Matt: Well first we’d just use our investor’s offices. So at the time Blacksmith, which was the predecessor to True Ventures, had a space in the Presidio. So Tony and I would just get together at restaurants and coffee shops and then, if we ever needed to have a meeting or something like that, we’d go to this awesome office in the Presidio. 

Later True, when they formed, moved to this awesome pier called Pier 38. They very presciently were like, “Okay, we’re a VC, we’re only this many people, but we can get this office space that’s a few thousand square feet. Let’s just put a few tables at it.” So one table was GigaOm, one table was Automattic. Upstairs was Bourbon, which later became Instagram. So there was a cool center of early startups that would have a little desk here.

Tony and I would go in there and meet or work from there but there was no reason to try to move the rest of the company there, because they were perfectly happy being where they were. I think Donncha was just about or had just gotten married. Immigration seemed really tough; there were a lot of barriers. So I was like, “Well, this is working, so let’s just continue to hire people wherever they are.”

However at the time it wasn’t clear that that was the thing we were always going to do. A lot of investors said, “This will work when you’re 10 people or 15 people, but at some point you need to get an office or you need to bring everyone to one place.” So we always kept that in the back of our mind as something that we might need to do. And [we] always had this small headquarters, first a desk, in the True Ventures space. And then we moved across the hall to a bigger space where we hosted lots of events and meetups. 

This idea that we needed something to be the official address of the company persisted really until… gosh, when did we shut down our office? Was that 2016?

Mark: It was a couple years ago, mhm.

Matt: We actually continued having a headquarters really until then. Partly this idea that on average, as we started to scale, about ten percent of the people we hired were in the Bay Area, and the ones who did like to go into the office in San Francisco. We assumed as we got bigger and bigger that the ratio would continue at 10 to 15 percent. And it turns out it mostly did. 

But what changed was that traffic and everything got so bad, parking got so bad in San Francisco, that even the people we’d hire in the Bay Area didn’t really want to commute. And then the tools for collaboration got better and better and so they didn’t really need to.

Mark: It’s interesting because there is a lot of this early Automattic history that you’re in San Francisco and you need to be in San Francisco physically in terms of the existing power and money, I guess, to be able to get a foothold for the company.

Matt: I think I felt like that as the leaders of the company we needed to be in San Francisco but the rest of the company didn’t. So it was great for Tony and I. We actually didn’t really have any new funders until 2013, 2014. And we were relatively unique at the time in that we had one West Coast VC, Blacksmith, later True, but our other investor was on the East Coast, called Polaris. So we were already distributed from our initial funding and that was unusual for an East Coast VC at the time, to invest outside of their geography.

Mark: Will we ever get an office back in San Francisco?

Matt: Yes, we probably will. [laughs] If something really interesting comes back up we’ll get a small space but it will probably be a secret. It will probably be just for board meetings and investor stuff, not something that is really an office like we have had in the past.

Mark: I want to switch over to some of the culture inside of Automattic since the early days. Automattic has its own creed. For those of you who have not read it before, you can go to Automattic.com/creed. But one key line from that is “I will communicate as much as possible because it’s the oxygen of a distributed company.” 

I know how it is today, and we can talk about that in a second, but in the early days did that mean that every single conversation was public, and there were no private conversations? What was transparency and communication like in those early days?

Matt: Yeah it really was, because there wasn’t really any reason to not have everything be open. The whole company was in one chat room. We had no teams, we had no managers, it was totally flat. We just organized around projects and code really. And there was such an advantage to knowing everything else that was going on. And we were doing enough things that you could know everything else that was going on and there was a lot of information sharing, or seeing the flow of code commits and code changes and conversations.

You got a shared collective intelligence of everyone at the company pitching in on every idea. So although we didn’t have formal product managers or a ton of designers or things like that, because everyone was using everything. We use WordPress ourselves, and seeing things that went through, you got a lot of real time feedback, almost like real time user testing with your colleagues, that ended up speeding up the iteration cycle and leading to some very usable products, even in the early iterations. 

Mark: That probably brings us to P2, [which] I believe was introduced in 2008. So for those who don’t know, P2 is essentially a WordPress theme for groups to share status updates with each other. I personally still see it as the foundational tool of all communication inside of Automattic to this day. But I went back and looked up your first post on Update-A-Matic, which is the all-company blog inside of Automattic, and I can see that you wrote the quote “Finishing up meetings.” That was the entirety of your P2 post.

Matt: [laughs] Well it was pre-P2. We just had an internal blog and we used it much like Twitter, which at the time was called TWTTR. The idea was just that — originally the prompt on Twitter, which is “What are you doing?” or “What’s going on?” or something like that. And so you would post literally whatever you were doing at that second. [laughs] So, finishing up meetings… 

Mark: This is like people tweeting what they had for breakfast or whatever?

Matt: That was the goal, actually. So a way for us to put things on there. And then we started to see that it was useful to have a place that you could have comments on or could do threads. We used it for keeping data. 

Mark: So P2 was not introduced yet as an actual thing, this was still just the internal blog?

Matt: It didn’t come until many years later. So it was really just this internal blog in IRC. What was, from the very beginning of Automattic, was this concept of meetups. Actually early on we did two ,I think, in one of those early years with five to 15 people. 

One that stands out in my memory, because it was an utter disaster, [laughs] was Stinson Beach, which is a place where I think we all fit in one, maybe two homes. We went on Stinson Beach, which is north of San Francisco, and we didn’t really know what to do at a meetup. So I think we assumed that we should just have all day long meetings and debates. [laughs] This drove, I won’t name who, but it drove someone crazy and it ended up being — I think there was a walk out of one or two people. [laughs] Voices may have been raised. It did not go well. 

At the time the valid feedback was, “Why are we sitting around, just arguing about things for hours and hours, not getting anywhere, not driving to decisions? And, by the way, we could do some of this written. Why does it need to be in real time, people reacting to information versus considering it, thinking about it and writing a response in a deliberate way?” Which — if you think about it — is a much, much better way to have a debate or an exchange of ideas than people just reacting in real time to whatever information is popping up in their head.

So that first meetup was a little bit of a disaster [laughs] and we decided to move them to once a year [laughs] and make them more organized. So we introduced the concept of meetup projects, or hack weeks. So the idea is we would get together and start and launch a number of things during that week as a way to do things we wouldn’t normally do, and just have fun. It was very much getting together and just working intensely alongside someone. [Which] was, and honestly still is really fun, [but] was hard to replicate with the tools at the time.

Mark: I think that holds to this day, this idea that the meetup should be a social event as well because you’ve got the entire rest of the year, and you have all of these tools in which you can organize and communicate and brainstorm. And so really pinpointing the specific things that you can only do in person together — and that’s also building relationships and building trust with each other — that you can then go back and be honest on Slack and P2. 

Matt: Exactly. Those meetup projects ended up creating a lot of things that became very crucial to Automattic today. So it evolved what first was called Prologue, which was the early version of a theme for WordPress, which put the posting on the home screen, so again, very much like a Twitter built on WordPress that looked a lot like what early Twitter looked like at the time, which was just a box and a list of posts.

And then that evolved at a later meetup into what we call P2, which basically took Prologue and made it real-time, so that as things were posted they would show up without you having to reload the page to get new comments or anything. So that was the evolution of what became what was and still is to this day the most important communication tool inside Automattic.

Mark: It’s an interesting segue to today because you’ve got P2 that are now hundreds of P2s, team and project P2 websites across the entire company. And then we have this global search tool in which people can search across every one of these websites. We also have a field guide which is more evergreen pages of documentation. 

But the P2 is where most of the deep communication is. And I’m raising this as a segue to today because seeing your original “Finishing up meetings” post and then seeing a post just the other day from Nick Momrik on our HR team, about the growing word counts of our P2 posts [laughter] as a trend, I think we can see that P2 has not fundamentally changed but how we use it has changed, especially once we started to embrace something like Slack.

Matt: Let me paint a picture of P2 so people listening can know what it’s like. So imagine on, kind of like a website, with a main area and a sidebar, and in the main area there is a post box at the top so you can type in there and click “post” and it shows up immediately, much like a Twitter or a Facebook. Below that is something that can be a post of any length but all the comments are in-line, so all the conversation happening next to a post, which again, doesn’t need a title, can be right there, and they pop up in real time. 

And on the right, on the sidebar, can be widgets that allow you to search, might have a list of people on the team, could have a schedule, could have notes, could have links to things, to resources, could have target launch dates or a countdown timer. And so the sidebar — it becomes kind of like your ultra-customizable home page for your team or project. 

The key for us was, as you mentioned — transparency was our default for most of our communication. The one thing that didn’t really work for that was email. So we’d email each other ideas and threads and brainstorms but then if someone new joined the team they had no way to catch up to those emails. And there was all this intelligence and data and wisdom being lost in people’s inboxes. 

Also, our email was busy at the time because support was all done through email and “support@” was just an alias for “everyone@.” So every support thread we got would go to everyone in the company and whoever replied first would claim that email or that ticket. 

So P2, Prologue, this internal blog, became a way for us to eliminate email. So anything that I would normally email, let’s say “Hey Mark, I’ve got this idea for a podcast, I think it should be X, Y, Z.” I would just post that as a blog post and you could see it whenever, and you still got that general intelligence that anyone else could see it and choose to interact with it or add their two cents to it, which is both a blessing and a curse. [laughs] It’s a blessing when you want the two cents and it’s a curse when you’re crushed under a sea of pennies. 

Mark: I want to go back to today. One could argue, at least from my experience — I joined, just for the record, in 2014 — but one could argue the past two years at Automattic have seen some of the most dramatic changes since you’ve started organizationally. We’ve grown very quickly. We’re now at 900 employees. And where we once were a strictly flat organization — we made up our own titles — we now have executive roles, product roles, and actual hierarchy. Can you talk about what’s changing and why at this stage and why it’s important?

Matt: Sure. A common misperception about Automattic was that we were non-traditional infrastructure. And that probably stems [from] — I avoided creating any hierarchy or really teams or normal company structure until we were maybe 50 or 60 people, and it became really, really necessary. So that’s where that comes from. But really since then, which is, by the way, the time chronicled in Scott Berkun’s book “A Year Without Pants,” which was — he was joining partially because I was like “Oh we need someone who’s done this before to help us create these teams,” but ever since then we’ve had a completely normal org chart, a completely normal hierarchy.

Although I’m a strong believer in how we work being non-traditional, distributed, it’s still really important for lines of accountability and for people to know where they fit and have one-on-ones and all those sorts of things. So that part of Automattic has always looked kind of normal. 

Mark: Essentially we’ve got team structures, which are — teams can be, what, eight to 10 people per team and they’ve got a specific focus? What is the ideal team size?

Matt: How we have approached Automattic is in a fractal way. So the idea is that if you zoom in or zoom out of Automattic it looks like the structure that is self-similar in many ways. So when Automattic was ten people, we had a designer, a businessperson, a bunch of engineers and we’d all work together to iterate on our area. 

So our team structure inside Automattic is very much modeled on that cross-functional idea where you get everyone together and you just — it’s that classic idea, a two-pizza team or whatever you want to call it. Get people working together with the autonomy and all the know-how that they need to ship and iterate with users, ideally as frequently as possible. Then when that gets too large, much like cell division, you split it into two identical teams and now you have two teams of, let’s say five doing cross functional work. 

As that grew, we eventually had divisions and now business units that are made up of divisions. Our largest business unit probably has two hundred and sixty people in both product and engineering and design. But still, if you zoomed in on one of those teams it would look very much like the teams that we had a decade ago.

Mark: And the tools too look very similar. So we still have the P2, as we’ve gone over, we use Zoom for video conferencing, as you and I are doing right here, then of course Slack came into the picture. That was shortly after I joined the company and that very much started with you, right? I feel like we were experimenting with it and then you decided let’s go all in.

Matt: I appreciate the credit. [laughter] With many things at Automattic, we give the teams a lot of autonomy. So we tried to adopt HipChat before other things that were a little better than IRC. It just didn’t take. I want to say probably the mobile team or there was some team inside Automattic that adopted Slack and then it was optional for people to get on it. Many, many did and we utilized their free version. 

I also knew Stewart from his Flickr days and so I knew a lot of the people involved with it. And in fact, a very early Automattician joined Slack pretty early on. I think she was one of the first 20 or 30 there. So we had that connection as well.

It just got so good. And my decree was more that we still had a chunk of the company that was really holding out on IRC and didn’t want to sign into Slack. And for communication tools it’s really, really important to have everyone on the same thing. And we were going through a lot of growing pains. Our non-IRC was mostly through Skype, or if IRC would go down we’d use Skype and that was just really awkward because then you would have to add everyone who joined the company manually. So it was just a very weird process. 

And so Slack, everyone being on there, the directory, all the things that everyone knows and loves about it, it was just so darn convenient. It really seemed like a better version of IRC. And, in fact, a lot of the conventions in Slack are directly modeled after IRC, like the reason why public channels have a hash in front of it. That is taken directly from how most IRC clients work. So it was pretty natural and, just for communication standardization, seemed important. So that was what I think eventually got our Systems team onto Slack.

Mark: I think it fundamentally changed how we even used our other tools, like P2. Whereas P2 maybe was more of a conversational tool in the past, now it’s a little more announcement-focused or this is our plan-focused [tool]. What would your take be on how P2 has evolved in the world of Slack?

Matt: It totally depends on the team. I would say this is one of the things that has changed most. I don’t know if it’s for the better but I’ll observe it in a neutral way. That as we hired more and more people that then come from the online collaboration or open source space — which, to be honest, there’s only so many people who have worked in an open source project or things like that. [laughs] They brought an approach to P2 that was a little bit more like announcements, like you said, and more meeting-centric, so needing more of that kind of synchronous real time communication and get everyone on the same page. 

And I think that coincided a bit with Zoom, which honestly makes meetings a lot easier than they’ve ever been and more pleasant than anything we’ve done in the past. So I think also people just wanted to connect better. We have a lot of psychological diversity in the company. I would say early on [we were] very much composed of introverts, including myself, for whom text was really our first choice of communication. And as the company grew, a lot more extroverts, or people who wanted to use voice or see each other as they converse. And so it was hugely controversial early on, even the idea that we’d have an audio-only meeting was widely debated for why that’s needed, if for nothing else than it’d be inconvenient for everyone not in a couple time zones. 

But teams can choose their own way. So as more and more teams had more and more people that were maybe composed of these or that wanted the real time synchronous communication and the tools got better, we started to use that more. And so on a team-to-team basis it varies a lot. 

I love, for example, our VIP team — that’s the enterprise part of Automattic — uses P2 I think the best of any team or division within the company. They put really everything through it. Their Slack is still busy but they really put a lot of thought and their P2 is great to read. It’s funny, it has GIFs, it’s fast, everyone has it pretty dialed in there.

Other teams have gone to where they might only make a few P2 posts a month, which if you’re doing that you’re not using it for daily communication, you’re not using it for real-time saying what you’re doing and keeping people up to date. You’re using it more like one-way announcements.

Mark: This raises the question of what skills are necessary to succeed inside of Automattic? Because I think there is a little bit of a — with P2 being so central and with written communication being so central and the fact that WordPress.com is a website and a blog hosting platform — that a lot of the most successful elements come around being able to blog internally as well. Do you believe that’s true?

Matt: I try to be pretty active on P2s. Call it an average of maybe a hundred posted comments per month. And my total word counts, I think it’s over a million words now that have been posted to these internal sites. And it’s really one of the richest treasures I have in my time in Automattic because everything is archived, everything is searchable, everything is there. And if I need to remember what I was thinking in 2012 when we made X, Y, Z decision, hopefully that decision is documented and the debate around it, and it becomes this huge source of wisdom, which I think allows the company to evolve in a more informed way. We try to only make new mistakes in having that entire history of the discussions and the collaboration that led to where we are today on these internal blogs. I think it allows us to move faster, smarter and better as we blaze a new trail.

Mark: It raises some other questions about talent. And one of the great promises of distributed work has been the idea that we can find talented people all over the world. But there are also trade-offs to what skills and talents this culture prizes versus, say, an in-office culture that’s more extrovert-focused or verbally-focused. What would you say the major trade-offs are and are they worth it? 

Matt: As I said, different teams work different ways. So there are teams that do very little written communication and communicate mostly through Zoom or audio or things. And so a nice thing about Automattic is whatever your work style is, you can generally find a team that matches that. If you want to be on a team that uses IRC and then basically never has any Zoom meetings, we got that. If you want to be on the team that has daily stand-ups, we got that too. And so by switching teams or divisions within Automattic you can get what might — you might need at a different company to switch companies to find your ideal work and collaboration style.

Q: You have mentioned here if a team wants to experiment with something — what are the experiments that you have maybe heard about recently that we haven’t tried inside of Automattic that you think we really should prioritize and try?

Matt: One that has come up is Invision, which is a distributed company, actually has everyone work East Coast hours. I don’t think that would be right for our entire company. We’re already too distributed for that, but it would be interesting if a team which was largely American-based agreed to overlap the exact same hours, if they would find that beneficial or just inconvenient. 

So that’s how I would want to approach any of these experiments is, say, find a team, at least 10 people, 15 people, who want to try something out, and see how it goes and have them do it for a fixed time period. Do a pre-mortem, do a post-mortem, see what are the learnings from it, and then what are the next actions we want to take from those learnings. 

Q: Is there enough creative tension or friction in a distributed company that maybe some form of self-censorship takes over, or we’re too polite with each other and we don’t raise the questions we need to raise? Do you think that is a valid concern?

Matt: This is why I have become relatively recently obsessed with the idea of an idea meritocracy. Have you read the book “Principles” by Ray Dalio yet? It’s a really impactful one in that you’re not going to agree with it but it will make you think differently. And one of the things he emphasizes is even in their in-office culture that the comfort that people have with challenging ideas and the openness people have to their ideas being challenged is crucial to getting the best outcome. 

Because if either of those is missing you get sub par outcomes. And because when you’re communicating on text it can be so easily misread, people hold back. I do think that’s 100 percent true. It’s less bad on audio or video but definitely in text communication you just don’t have that nuance. All the emoji in the world can’t recreate the kind of timbre and tone of voice and all the additional data we get when we’re actually talking or seeing each other. 

I think a lot about that, and I’ve been thinking a lot about how to foster that. Google talks about psychological safety being key to the high performance of teams. How to foster that where we think that good ideas can come from anywhere and everyone is comfortable presenting and defending vigorously anything they believe in.

Mark: One other thing I’d like to do real quick is just a quick speed round in equipment breakdown. Tell me a little bit about your workspace. Now you travel a lot; you’re in a lot of different cities, tell me a little bit about your must-have items?

Matt: [laughs] Yeah, check out the What’s In My Bag post for the kinds of things I use every single day. It’s the 15-inch MacBook, it’s the cables, it’s the Sennheiser headset for being on calls. I find audio quality is far more important than video quality for creating a great meeting. At home to me that’s the ultimately luxury. [laughs] 

My favorite part from being distributed is when I am able to work from home, which is probably a minority of the year, but I love it because you can have your music you like. I like having a candle on the desk. [laughs] I like the temperature to be a little bit warmer because when I work my hands and feet tend to get extremely cold for some reason. Normally they’re fine but for some reason when I’m on the computer they get ultra cold. I recently installed the instant hot water thing at home. So in my sink I can get near boiling water and make tea. On my desk I always have a notebook. I love paper notebooks for writing things down, I find it’s a lot less distracting than trying to type things out. 

For a while I used to try to have desktops, like an iMac, and in fact the new iMac Pro is — I still have one of those in Houston. But I’ve really gone to where I like a great monitor, [and plugging] it into the 15-inch is the easiest. And I love these new ultra wide monitors. LG makes them. I think I have one that’s 34-inch and one that’s ginormous, 38 inches wide and curved. And using that, it feels like a panoramic experience. 

I knew that I was spending too much time on Google Sheets when I was like, “Oh my goodness” and I stretched out the sheet all the way to the 38 inches of this ultra wide monitor. I was like, “I can see everything!” It felt almost omniscient. But I was like, “I’m in spreadsheets too much.” [laughs] I need to load up some JavaScript.

When I’m at home, I use a Logitech BRIO camera, which is a high-end, 4K-webcam that they have. I find it has a much better aperture so it creates better colors and light in low-light situations, which is often where it is. If I don’t have good natural light wherever I am, I have some desk lamps — just soft light that I can put there so I don’t look weird or back-lit [laughs] wherever I am. If I’m on video I try to — I actually think of it probably much like a webcam YouTuber would. Like, “What’s the lighting like, what’s the audio, how can I present well there in a professional manner?” And I actually see a lot of folks at Automattic curate their background, having things they like in their background. Because you actually end up looking at yourself, what’s behind you, a fair amount when you’re on Zoom or Hangouts or one of these things.

Mark: A couple quick speed round items. Phone calls, love them or hate them?

Matt: It’s actually a resolution of mine to pick up the phone more. I don’t really receive phone calls because it’s all spam now but I love all the non-phone things, like Facetime audio, Zoom, Slack Audio is actually really good. And one thing I’m trying to do more is actually switch mediums. So if I’m finding I’m having a long text conversation with folks, I try to balance the audio. 

And it’s also nice in that some folks — my mom is getting older and sometimes it’s hard for her to read the text. So it’s always really nice to hop onto that. Video is really nice too. I got my mom one of these Google home devices that had a built-in camera. So you can use Google Duo — that’s actually the only reason you would ever use Google Duo to make a video call — to her and it lights up this device that’s by where she usually hangs out. So I’ve found that’s actually been really, really lovely as a way to drop in and stay connected with a loved one who I don’t see as often as I would like. 

Mark: Final question. Twenty years from now, what percentage of jobs will be distributed?

Matt: You’re turning my questions back on me. I am going to say… Well, I kind of don’t want to say so that future guests can answer this without knowing my answer. I do believe there is a window where distributed companies have a real advantage for recruiting, retention, and everything. That window is probably three to five years before the incumbents really embrace this. 

I then think that job seekers are going to learn to ask more sophisticated questions. So they won’t just say can I work from home or not, they really dive in to where do decisions get made, where the center of gravity for the organization is, can their career advance as much being not where the headquarters is or in a distributed fashion as it would if they were say in Mountain View, at Google. 

And those types of more subtle questions will be the things that — as the internet giants embrace the surface of distributed work but perhaps not the deep spirit — all the things that are currently bundled with that — and startups like Automattic or Invision or UpWork — I guess are we still a startup? But companies like that. The bar will change. I’m looking forward to that happening.

Mark: Thank you, Matt.

Matt: It’s been fun chatting. 

Mark: That was Matt Mullenweg. Thanks again, Matt, for taking the time to speak with us about the past, present and future of Automattic. If you want to read more from Matt, you can always go to his blog ma.tt, and he is also Photomatt on Twitter. Thanks for listening.