How WordPress and Tumblr are keeping the internet weird49 min read
Matt Mullenweg is the CEO of Automattic, the company that owns WordPress.com, which he co-founded, and Tumblr, the irrepressible social network it acquired from the wreckage of AOL, Yahoo, and Verizon.
Matt’s point of view is that the world is better off when the web is open and fun, and Automattic builds and acquires products that help that goal along. That bet is perhaps most pronounced with WordPress itself. Around 43 percent of all the websites on the internet run on WordPress, which is a piece of open-source software anyone can download and use for free. It is officially administered by the WordPress Foundation, but if you don’t want the hassle, you can just go to WordPress.com and pay Automattic to do most of the work for you. It is an absolutely fascinating model that has obviously worked really well, and I wanted to know more about why Matt set it up that way and what the challenges of that structure are.
I also wanted to talk about Tumblr — it’s Taylor Swift’s favorite social platform, and it is one of those things that users have kept alive no matter how many corporate owners have tried to kill it. But like every social platform, it has meaningful moderation challenges. Famously, to get an app on Apple’s app store, it had to ban porn, which users are still upset about. Matt and I talked about those challenges.
One note: Matt mentions something called “Gutenberg” a few times — that’s the new creation experience in WordPress, the part of the site where people actually make things.
This is a long one, but it’s deep. I think you’ll like it.
This transcript has been lightly edited for clarity.
Matt Mullenweg is the CEO of Automattic, the co-founder of WordPress, and the CEO of Tumblr. Welcome to Decoder.
Very happy to be here.
The last time you and I chatted was right after the Tumblr acquisition. That was a fascinating conversation, but it was years ago, so there’s lots to follow up on. Automattic has bought a lot of companies. You released a lot of products — started with WordPress, which you obviously founded. You acquired Tumblr, bought Pocket Casts, you have WooCommerce. Start at the beginning: What is Automattic? How do you think about it? What are all of your companies?
Sure. So Automattic has been a fully distributed company since 2005. For 17 years now, we’ve been trying to make the web a better place. We want to democratize publishing in commerce. We’ve been saying that for a long time, before democratizing was cool.
To do that, we both build, buy, or partner with things that make the web more open and fun. Pocket Casts is one of the very best podcasting apps. It’s very user-centric; it’s an independent alternative to the Spotifys and Apples of the world. We’re just relentlessly iterating on it to try to respond to user requests and make it better.
Most of our business models are through people paying us, as opposed to advertising or other models. We provide upgrades and that recurring revenue is what allows us to come to work the next day. We’re about to come up on 2,000 people working full time with Automattic. It’s really grown a lot, even since the last time we talked. I think we hired over 700 last year.
700 people. Where did you hire them all?
All over the world. We have folks in 93 countries now. From the very beginning, and especially last year, one of our big advantages is that we’re not really geographically restricted in where we hire people. We are strong believers that talent is evenly distributed in the world, but opportunity is not. When we provide opportunity to places where they don’t normally have it, we find really, really amazing people and they tend to be great colleagues that stay for a really long time.
The other unusual thing we do is we pay global pay rates, which also helps for international recruiting, so we pay similar ranges for salaries, regardless of what country you’re in.
That sounds great. I just want to dive into the practice of it real quick. You want to pay someone in your 92nd country and you only have one person there. Your compliance cost to do that is very high — you have to register with a government and figure out their tax situation. How do you manage all that? That’s a lot of overhead to be that distributed.
Yeah. There are companies as well that can manage this for you. Even in the US, when we started, all 50 states have different rules and different counties have different rules and everything. Most people end up working with a payroll provider. There are some international equivalents.
The other answer, just very simply, is that some of these people are technically contractors who bill a fixed amount every month. That removes a lot of overhead, especially for places where you might just have a single person in a country. It puts a little burden on folks to have to handle their own taxes, and they usually register a corporation for contracts and things like that, but it does simplify things a lot.
I’ve always loved this model. For a lot of people who’ve tried it, it seems like there’s actually some amount of complexity you have to manage through. In practice — let’s say I want to find the best podcast host in the entire world. How would I do that? If there’s one person who’s the best at it for my role who is in Luxembourg, how am I going to find that person? How do you go find these people?
Something I think companies don’t do enough is just promote to their user base or promote to their audience. There’s probably three roles that you’re really trying to hire for Verge right now. Say those. Run an ad for them. You run ads for other people — run ads for yourself.
We try to put Easter eggs all over our products that lead people to our jobs page. We’ve tried to make all the “about” pages of our apps a lot better, like Day One, Pocket Cast, Tumblr. If people are in there, point them to your hiring. Your existing audience is a fantastic way to do that, especially if you’re like us and have an audience of hundreds of millions of people.
You started to mention some other products: Day One, Tumblr, Pocket Casts, WordPress. What’s the suite? Can you list them all?
Sure. I’ll say the big ones: WordPress.com, WooCommerce, and our enterprise business called WordPress VIP are the drivers of Automattic’s business. A lot of people think of WordPress.com as a simpler place to host WordPress, but they don’t know that it actually can run all plug-ins and themes for WordPress and it’s all auto-updated and auto-secure. Think of it as a place where you can run any WordPress site — it’s totally bulletproof, highly secure, updated and can get basically as much traffic as you can throw at it.
WooCommerce is open-source Shopify. I think it is the only e-commerce platform growing as fast as Shopify. We did $31 billion of goods sold through WooCommerce last year, which I think was doubling year-on-year. You can sell pretty much anything online and it integrates with WordPress.
Enterprise has been fun. It’s basically like large enterprises who want to use WordPress — hopefully, maybe The Verge someday. Some other sites use it like TechCrunch. One that just switched over is whitehouse.gov, which is very interesting right now with all the things happening internationally. We can host any size of WordPress that you can imagine and also make it very, very secure. Basically, it’s like the beefed-up version of what we do for all of our users.
That’s what drives the business. I don’t want to distract from these other things, but we also have always had what we call “five for the future.” This is an idea in the WordPress community that companies put 5% of the resources back into open-source. This basically avoids the tragedy of the commons. Five percent of the 2,000 employees work full-time contributing to open-source. They’re not working on any of our commercial products or anything like that. I feel like that’s honestly the bare minimum that we can contribute back to make the community better.
We also commit 5% to new stuff — our version of other bets or experiments or labs. That’s what these other products fall into, like Day One, Pocket Casts, et cetera. These are typically really well-loved products that have a very passionate user base. Simplenote is another one that’s really popular, and The Verge has covered it really positively in the past. We want to cross-promote that and bring our long term investment to bear to see what could happen. That’s what WooCommerce was seven years ago; it was kind of small and wasn’t making a ton of money. Now it’s a powerhouse. It’s going to be as big as the rest of the company combined pretty soon. Some of these other products, including Day One, have that potential.
You mentioned the tragedy of the commons. That’s a powerful phrase. What do you mean by that, specifically as it relates to open-source software?
Tragedy of the commons is from economics actually. It’s a story. There’s a common field that belongs to this town, but it doesn’t belong to any one person. If all the farmers brought their sheep and cattle to graze in that field, but none of them were investing in maintaining it — maybe not having their particular sheep or cattle lay off it so things can regrow. The field gets overgrazed and dies. No more grass. Everyone loses.
In open-source, it’s very easy for companies to use open-source without contributing anything back, but that’s kind of one of the features of it. We can’t complain about it really, because that is what the license says you can and should do. But I think that companies who think more long-term say, “Okay, I’m getting a ton of value for this. I’m not paying a penny. How do I make sure that this is around five or 10 years from now?” We’ve seen examples of libraries that the whole internet depends on.
Yes. This is where I was headed.
We also have examples of open-source projects that reach an exit velocity. They become this positive flywheel — more people using them means more people contributing to them. They become just totally ubiquitous, almost like natural monopolies, but not evil like monopolies — monopolies that belong to everyone so they do good. That could be things like the Chromium browser engine, which is now powering Internet Explorer, Microsoft’s product, in addition to Brave and Chrome. WordPress in the CMS space now has 43% of all websites. It’s growing faster than all the others combined. That will reach probably 80–85% in the next decade.
Wait, you just predicted that WordPress — both the open-source and WordPress.com, I’m assuming — will power 85% of websites in the next decade.
I think that’s what we could do. If you look at the number of handsets powered by Android, it’s about 85%. I think Chromium as an engine is going to get to 85% or 90% browser share. You’ll always have 10 or 15% of other stuff, new stuff, et cetera. What happens is when you get the opposite of the tragedy of the commons — when you get the abundance of the commons, it becomes a super future. Like I said, this positive flywheel where the more people that use something, more people that contribute, the better it gets.
Since it’s totally free and belongs to everyone, there are more reasons for people to use it. The software can evolve really rapidly. If you think of the rate of evolution, it just outmatches anything, including proprietary competitors that might have hundreds of millions of billions of dollars behind them. They just can’t match the brilliance of the world working together on a single thing.
You said Automattic is 2,000 people now. You hired 700 new people in the last year. How is that structured? Who works where?
We call ourselves a digital Berkshire Hathaway. Myself as the CEO of the company and some centralized services — including HR, legal, systems — try to serve these relatively independent teams that have a lot of autonomy for how to organize, run themselves, and iterate on their product. This middle part of the company, call it 200 people, is doing everything we can to remove barriers and help the rest of the company go faster.
The teams run like little mini companies. They have their own CEOs, executives, and CMOs. They’re developing their own boards actually. Part of what the centralized part does as well is say, “Hey, we’re doing this same thing in two places. We’re building a newsletter thing over here for WooCommerce and we’re building a newsletter thing for WordPress.com. How do we combine that and make sure that they work with each other and they complement each other?”
All org structures are a series of tradeoffs. So we tradeoff for speed and autonomy, but then sometimes we have to come back and make sure we’re working with ourselves pretty well.
The secret of Decoder is that, fundamentally, this is a show about org charts. I don’t know if anybody warned you.
Somewhere inside of the show is a series of org chart conversations every single time. You’d be surprised.
I’ll throw this theory at you: the most revealing question you can ask a CEO is, “How did you structure your company?” Everyone has to answer that question because they need to know. If you don’t know the answer to that question, you’ve revealed something very important about yourself as a CEO. Whereas if I just came and hammered you on why you didn’t solve this moderation crisis, the answer might be revealing, it might not, but it’s not the central truth of how you organize a company. That’s my thesis. I’m curious what you think about that.
It reminds me of a quote by, I think, Picasso: “When art critics get together, they talk about the philosophy of art and what this or that movement means — but when artists get together, they talk about where to buy cheap turpentine and what brushes they’re using.” The conversation becomes a lot more tactical, but that’s actually where a lot of that artistry comes from.
It is true that when I get together with other executives, we often obsess about organizational design, but what I’ve learned there is that there’s no one design that is better than another. You’re choosing a series of tradeoffs and you need to be deliberate about what trade-offs you’re choosing. One example of a tradeoff we’ve switched on over the years: For a while, our design was totally distributed across all the teams, because we want these teams to be cross-functional.
Product design, yeah. For a few years, we actually centralized design. We brought all the designers into a central team reporting structure. We had a really awesome strong lead there. That was part of improving the quality of what we’re doing, but once we accomplished that, we put the designers back out into all the teams and divisions.
The other fun thing I like to say about Automattic is that it’s fractal. When you zoom in or out, it’s self-similar. When the entire company was 20 people, it looked a lot like what a team of 20 people looks like now. We try to make it so there’s a natural growth and division of teams.
That’s fascinating. You described three things as all the business and a bunch of other things as other bets. When I look at other companies that are organized like that, there’s a natural tension — when are the other bets going to become bets? How are you thinking about that? How are you determining when Pocket Casts has graduated out of other bets and become a real business?
An easy threshold would be 10% of revenue for the company. If it’s under that, it probably is going to be part of the other bets structure and maybe report up to a common executive or be abstracted a little bit. If it’s above that, it’s a serious business for us and it’s going to have its own representation that people are going to talk about it to the board every quarter and things. I’ll talk about it with you.
The tricky part is that we have — and particularly the press has — an addiction to novelty. The new thing draws a lot of attention, but you have to balance that. We’ve launched like 30 new things on WordPress.com this year alone that drastically changed the user experience. Gutenberg is redefining not just WordPress but CMSs in general. We’re about to put it on Tumblr as well. There’s some really exciting stuff happening, but it doesn’t always necessarily fit easily into the narrative.
Don’t worry, this show is all about org charts and decision-making, and then we’ll get into WordPress. Actually, this is the Decoder question: How do you make decisions? When you’re deciding to centralize and then decentralize design, how do you come about that decision?
Me personally, or the organization?
Both, and how are they different?
There’s definitely folks in Automattic who want me, Matt, to have answers. They’ll ask me a question and be like, “What is the vision for X, Y, Z?” or something like that. But almost everything we’ve done in designing the company is saying that the answers are at the edge. The wisdom is in the people talking to customers, doing the work, writing the code, designing things, doing user tests. We want to make it so the decisions come from the people with the most information.
I try to make as few decisions as possible and really say, “How do we push that out to the teams, the divisions, the edges of the branches of the tree versus the trunk?” The assumption is that none of us are as smart as all of us, so now what do we build in there?
I ask for transparency, so that things are written down, shared, communicated. I love the idea of decision journals. We use this internal blogging system built on WordPress called P2.
I was going to say [WordPress competitor] Movable Type.
Yes, it’s built on Movable Type. [Laughs] What’s interesting at Automattic is there’s no internal email. I get a handful of emails a year from my colleagues. Everything happens on these internal blogs. What that means is we have essentially an organizational blockchain where every single decision going back to 2007 is on one of these internal blogs. You can find how every piece of code works, or every business decision, or every logo. Everything is in there somewhere.
Even if you and I decided something in a meeting, we need to write it up afterwards. It’s on this P2, so people can participate in it asynchronously. Future generations or future versions of ourselves who’ve forgotten why we made a decision can tell why we did that.
Finally, we try to say, “Reversible decisions quickly, and irreversible decisions deliberately, or slowly.” We put pretty much every decision into two categories. Most — 99% of what you do — is very reversible. Some things are really big. Who you take funding from, acquisitions — these things are hard to unwind, so you need to make those decisions very deliberately.
What’s a decision that comes to you that you have to make regularly?
What’s funny is — especially across 2,000 people — the things that come to me are the things that no one else has been able to resolve yet. Ideally, the problem made it through a lot of layers of really smart, talented people trying to resolve it before it got to me. That means that my job is never dull. I see the edge cases. I see the things which — across 2,000 people in 93 countries — are the stuff you never thought of. They’re completely novel.
Let’s talk about WordPress. It is a giant. It is perhaps an overlooked giant, as you were alluding to earlier. You did just roll out this big, new upgrade to it called Gutenberg. How big is WordPress? How many people work on it?
At the company, it’s probably about 500 people working on WordPress, but for every WordPress release, like version 5.9 that came out in December, we list all the contributors. People employed by Automattic are typically 10% or less of the contributors.
That’s one of the cool things about the WordPress community. Even though we’ve had this for-profit that I started 17 years ago, Automattic, the community has grown and thrived over that. There’s tens of thousands of plug-ins, themes, billions of dollars of business that goes through other companies and the ecosystem. That’s something we very deliberately work on.
When you study platforms, whether that’s Windows, other operating systems, platforms that were real platforms versus platforms that weren’t — there’s typically a ratio that comes in, which is 1:20. The creator or the base contributor of the platform, or their commercializer, or whatever you want to call it — if they’re making more than 5% of the revenue in that ecosystem, they’re probably suffocating the ecosystem. It’s interesting, also, if you apply this to app stores.
I would also argue that there’s a lot of money to be made on top of app stores. Think of every DoorDash delivery, or Uber ride taken. Even though the 30% cut seems egregious and probably unsustainable, if you were to look at the total ecosystem of value built on top of Android iOS, actually I would argue that they’ve been very successful platforms. Apple’s take of all the revenue is probably less than 5% of the value they’re creating in the world.
That’s what we target. It’s also what distinguishes the WordPress ecosystem from the proprietary competitors like Shopify or Works or Squarespace, which typically will make 50–95% of the revenue in their ecosystem.
I think you might be our first open-source CEO on the show, depending on when this runs. How do you manage that? WordPress is a big open-source project. Anybody can go download WordPress and start a site. You’ve got nothing to do with it. No revenue flows to you. Anybody can modify it and do whatever they want with it. Then there’s what they contribute back to the main open-source code so that you can release new versions. You guys manage that. Then you have a business next to it. How does all of that work? Is there one person who’s in charge of WordPress, and is approving the patches that get submitted? How are you managing that aspect, and how are you deciding what to make and release?
I was WordPress’s lead developer before Automattic started. That’s the first hat I wear. You dance with the one that brought you. When I’m thinking long-term, I’m thinking first and foremost, “What will be best for the WordPress community this year, 10 years, 30 years from now? What will make us the most sustainable, the most resilient, the most antifragile?” I think of that first.
I also run Automattic in a way that its economic self-interest is totally aligned with what I believe to be the best long-term for the broader WordPress community. That includes things like the way that we’ve chosen to monetize. For example, a lot of open-source projects create an extra features version when they monetize. That’s usually how they get you to upgrade. We’re very explicit: What you can buy from WordPress.com, you could also buy from Bluehost, or WP Engine, or something.
The challenge there is that you could actually buy WordPress from anyone. It’s literally a commodity in that anyone can sell you something called WordPress. By the way, we put all the best stuff into that, like Gutenberg, which vastly expands what you can do with WordPress. That’s the core. It’s free. Anyone can use it. In fact, that’s part of what’s so powerful about it. It becomes a new standard. We don’t hold that back, but sometimes we do need to charge for something. Where we draw that line, and how we draw that line, is very tricky. I call it the Robin Hood business model.
There’s this plug-in called Akismet. It keeps spam off your site. It’s the best anti-spam, web anti-spam system on the web. A lot of people don’t know, but it’s actually the first code I wrote after leaving my job.
The first commercial thing we created before WordPress.com was actually Akismet. It works really well. It’s free for personal use, and then paid for commercial use. It’s kind of an honor system. We don’t really police it that much. It’s a Robin Hood business model.
We provide this free service that keeps spam off 43% of the web, because almost every WordPress site uses it. A very, very, very small percentage of people pay that, but that small percentage of pay makes it sustainable. It makes it so that we’ve now been able to use, and develop, and battle the spammers, battle the bad guys now for 17 years.
When you are thinking about how things are going to make money — you have lots of competitors and things of commodity — what leaps to my mind is WooCommerce. WooCommerce is an open-source payment stack. It seems like websites in 2022 are mostly e-commerce sites. That’s mostly the web activity, it’s a handful of news sites and e-commerce sites. That’s the whole web. Wordle is the most unusual website in the past five years because it’s a web game. People went nuts, like, “This is the old web.”
Do you see something different happening in that split? It is fashionable to run around saying the web is dead and that apps shape the world, but in my mind, the web’s pretty healthy for at least two things: news and shopping.
I think that’s your bubble, if I’m totally honest. That’s what’s cool about the web: We can live in a bubble and that can seem like the whole thing. One thing I would explicitly try to do in 2022 is make the web weirder.
That’s great. I fully support that project.
It’s also Gutenberg for WordPress. We’re like, “Let’s just stop writing texts in a box. That’s boring. How do we give you full layout controls — the full ability to create something like, was it ‘Snow Fall’ — that New York Times story that blew everyone’s mind? How do we put that creativity—”
I want to say The Verge filed a “Snow Fall” story before the Times. Just my personal note.
Actually, 100%. When y’all launched — you still do it with your editorial features. It’s awesome.
We actually have a publication called Atavist that publishes only once a month. Every story is like a visual journey — like the best magazines, basically. We want to make that easy for every single one of your authors to do with a few clicks.
That’s part of also what is cool about a system like WordPress, is sometimes I describe it as Promethean. We take the fire from the gods and bring it to the people. You created this fire, this cool design layout. You probably have a proprietary CMS, but the only people who have access to that are people who work for your company, or that might license it out, or something like that.
How do we take these cool ideas and bring them to literally the whole world — where regardless of what language you speak, no matter if you have any money, we can give you these cool tools? It’s not just news. There’s a ton of personal sites out there, a ton of personal blogging, so much art, so many artists that are sharing. Beeple, who sold an NFT for $69 million, has published on Tumblr every single day. That’s where he published. That’s where he got started. That’s where he still posts every single day. I apologize, the name escapes me, but the biggest movie star in India is publishing to Tumblr every single day. He’s done like 4,000 days in a row, without fail.
There’s people that are sharing. That’s really, really exciting to me because it’s kind of like the personal web. It’s different, it’s unique. I go to a Medium article, or even a Substack article, and I read the article and I just remember Medium or Substack. I actually forget the author because the sites all kind of look the same. How do you bring that — not just the editorial voice, but also the visual personalization — that you can really have your own home on the web? You don’t just look like everything else.
I buy all of that. I guess “news” is too flat and too narrow of a description. There’s just not a lot of the weird web. The web that you would’ve thought of where I made a little web tool that mashes up a map with another map. That early Web2 spirit seems to have faded away into apps, or features of social networks.
I’m just curious, where do you see the growth for WordPress? Some weird stuff, we hope. Maybe a lot of artists are going to sell their NFTs, but I just think about where your money would come from. It seems like e-commerce is actually the place that you would be most focused.
Yeah. We’re doing a good job at democratizing publishing. WordPress is on the right path there. Like I said, I think it’ll get to an 85% share. I now feel so strongly about them doing the same thing for e-commerce, because I think that we need those same freedoms: freedom to publish, freedom to transact, freedom to use any payment system, freedom for the transaction fees to be just as low as humanly possible versus going up every year. We need open-source alternatives, not just to Shopify, but also Amazon, and Etsy, and everything else.
Those are still great services, by the way. I give Amazon a big portion of my paycheck. I love it. I actually would argue that it’s making the world a better place, but we still need an open-source alternative to serve as a check and balance in the free market towards these successful companies.
What does WooCommerce do for somebody? You have a snowboard store — that’s the famous Shopify origin story. Instead of having to build Shopify, you decide to use WooCommerce. How does that work?
It’s a plug-in for WordPress. You could go on your WordPress dashboard, search WooCommerce, one-click install it, and then you’d be taken through a workflow that’ll help you set up a merchant account, set up your first products, and connect it to shipping.
Commerce is so complex. You have taxes, shipping, inventory management, but because WordPress is open-source, there are a ton of extensions built for it. You can pick and choose and put things together to have your own store on the web. Even now, you can have a point of sale. If you have a physical store, you could have a WooCommerce little reader, integrate for Stripe, et cetera. People can tap to pay and that will synchronize with your online store. You can do some pretty fun stuff.
Actually, a store we just launched — kind of secretly for Tumblr’s 15th anniversary, is shop.tumblr.com. We’re putting Virgil Abloh-inspired, very limited drops of cool high-quality stuff. We were able to launch that on WooCommerce in about two weeks, which shows you can really do anything.
Now, why would you choose it versus one of these other ones? Shopify, by the way, is an awesome product and an awesome company. Toby’s an inspiring leader. I really like them a lot.
We had Harley on the show a few months back. I ended up being like, “Well, I’m going to go work at Shopify.” They’re big personalities.
I do also think that you’re starting to see at the edges the complaints of their ecosystem. They’re having so much of the revenue flow through them, and their commercial embedded growth obligations are driving them to conflict with their community more and more. The community is what really made them, I think. Basically, what we did with WordPress is what we’re doing now with e-commerce. WooCommerce is free. It’s open-source. At the low end, we just ruthlessly commoditize it.
The average person on Shopify is paying $1,200 per year. That’s really expensive. It doesn’t need to be that expensive. It can be one-tenth, and then someday one-hundredth of that, to have all that functionality. WooCommerce is one-tenth that today, and it’s going to get cheaper and cheaper every year.
When you say it’s open-source, can I just take it and use it for free?
But no one does that. Most people are paying you $120.
We try to make it easy to pay us.
WooCommerce is also different from WordPress in that WooCommerce is wholly owned and run by Automattic. It’s not primarily a larger community. It does have its own community, but it’s owned and run by Automattic.
The other reason people leave Shopify and go to WooCommerce is at the high end. We have this guardrail approach where we get to ruthlessly commoditize the low end, and on the high end we allow customization. Ultimately, a SaaS product will always hit a ceiling. You might run into something that is different about your business model versus the SaaS provider.
Open-source is true freedom, especially once your store starts to make a hundred million dollars plus per year — even small things that could be a half a percent, or a percent, on every transaction, start to really add up and be millions of dollars of value. That’s why very high-end stores switch to WooCommerce.
I’m super interested in that kind of competition. You’re not head-on Shopify 2, this is a different model and you can make a series of business and personal and philosophical choices that lead you to our model versus theirs, but then broadly you have some of the same challenges.
Apple turns off app tracking transparency, Facebook’s ad rates plummet because it’s harder to target customers. I get a lot of ads for spoons lately. I don’t know what’s going on there, but it’s hard to find me, the one person in America who’s in the market for spoons on Facebook. The little e-commerce store that has the one hot pink spoon I’m looking for doesn’t find me as cheaply. They fall apart. At the end of all that, Shopify stock prices crashed. You can see that Apple turning off ad tracking transparency had this big ripple for the e-commerce stack. Did that hit you?
WooCommerce stores who advertise on Facebook probably were impacted, but the reason why people go to WordPress, or WooCommerce, is to build a direct relationship with their customers and not be entirely dependent on or mediated by Google, Facebook, et cetera. So they build mailing lists. WordPress has the best SEO in the world and WooCommerce inherits that. These WooCommerce stores are findable on an organic search, not just paid search. And because they integrate so well with WordPress, they develop amazing blogs and people follow these blogs.
These are the things that make WooCommerce stores hopefully more resilient to these types of changes. Over time, as part of being developed primarily in an open-source way, we really err on the side of radical user privacy, radical ownership of data. We end up being on the right side of history when regulations or new things come in.
Where open-source gets in trouble is when the regulations are essentially written by the incumbents — something like a GDPR, which is written in a way that really benefited Google and Facebook and penalized every other person in the marketplace. That’s tricky when you have that sort of regulatory capture, but again, I always think about these things long-term. That could happen for a few years, but if you look out five, 10, 20 years, ultimately, consumers and economics will drive these decisions.
Let’s talk about Tumblr real quick. I think I have a handle on WordPress, which at this point is kind of like a B2B product. Some individuals are using it, but the thrust of it is that lots of businesses are using it. They’re paying you it sounds like whatever amount of money they want to pay you at whatever time. You’ve got a big commerce stack next to it. That feels pretty enterprise-y.
I’ll tell you a stat most people don’t realize. Half of all users who sign up for WordPress.com every day are there to blog.
That’s amazing. They’re still doing it.
To be honest, even internally, we assumed everyone was coming to us for CMS features, and I think we over-indexed on that more business-y side that you just described. That’s also because we thought more revenue was coming, but when we sliced the data differently, we actually found that more than half of signups were there primarily to blog. I think it’s cool that people are still blogging.
Yeah, I do too. We could end it here. My heart is warmed. No further questions. That’s great. I’m happy that’s still happening. Enterprise-ish is what I would call WordPress.
It is more enterprise-y. One caveat I’ll say to WooCommerce is that it’s a developer product right now. It’s like a Stripe or Twilio. It’s really developer-first.
Right next to that is Tumblr, which is the most consumer product that has ever existed.
It’s super consumer.
But it’s also super consumer in the sense that it’s a playground for artists and musicians. Taylor Swift is reading Tumblr, and that shows up in the lyrics. There’s a feedback loop between the lyrics and the Tumblr people. It’s nuts.
The things that happen on Tumblr are utterly unique to any other social platform. You bought it out of the wreckage of Yahoo in 2019. Back then I asked you, “What are you going to do with it?” You replied, “We’ve got some great ideas.” But the first thing you had to do was ban all the porn so you could get the app in the app stores.
That happened before we bought it. That was actually under Verizon.
That was under Verizon, but it was going to happen no matter what. If they hadn’t done it, you were going to have to do it.
I think Verizon is super conservative. If you look at apps like Twitter and Twitch, they actually do have an astounding amount of adult content, but they do it in a way that is compatible with Apple’s rules. So I think Verizon took an extremely conservative approach to try to get rid of even borderline things — things that I would call art — from the app. They did it in a way that penalized a lot of legitimate users.
One user contacted me after the acquisition and said, “I posted a picture of my manicure.” A picture of their hands — some algorithm thought it was too much skin tone. Their account got banned. Then because the support was so backlogged, no one responded to them for months.
What’s a good way to kill a social network? Ban your most active users, and don’t reply to them for months. People will go someplace else and find someplace else to publish.
But I’m just looking at our conversation from back then. You’re correct. Verizon had done it, but your position back then was that Verizon might have been more conservative, but you said to us, “If you want big policy changes here, put pressure on the app stores. No one has any leverage.”
Yes. Still true. Tumblr was in the news in December. We were scrambling right before Christmas because Apple came to us — you probably saw where we had to ban these random tags. The app store review process can be very inconsistent. Sometimes something you’ve done for years and years, they’ll come to you and say, “That’s not allowed anymore.” Or, “We searched for this search term and now your app is banned unless you make these drastic changes very quickly right before Christmas.” It’s very tricky.
That was really the heart of my question. You bought Tumblr in a moment of crazy back and forth with app stores. Tumblr has had a bit of a Renaissance in the pandemic. People are using it again. There have been glowing write-ups of it here and there, but it’s still in this weird spot because its brand is so consumer and so arty and wild, it runs into app store moderation policies maybe more than any other product that I can think of that is like it.
I think we talk about it more than other products, which is maybe why we get into trouble more. Maybe it’s retaliatory. If you have drinks with anyone who works on this process at pretty much any large app-based business, including Twitter, they’ll tell you the same thing. Apps can get blocked randomly one day, things you never even thought of, and your team’s scrambling to address it.
There’s been a lot of noise. There have been hearings. We have had David Heinemeier Hansson ranting and raving on our shows. There’s been a lot of scrutiny of Apple. There are potentially some bills in Congress now. Is the posture changed at Apple and Google about this stuff or is it still the same as it ever was?
Hmm. You don’t hear about this stuff with Google very much. So I will say that the Play Store is a lot more consistent in their application of things. Honestly, Apple has a tough job here, so it’s not dissimilar to trying to moderate user-generated content. It’s that they’re moderating hundreds of thousands of apps and the team that does that — I have no idea how big it is, but let’s call it at least 1,000 people. I literally have no information. That’s just a guess.
You’re going to get new people. You’re going to get mistakes. That’s just part of it. I do appreciate that we’ve been able to resolve everything so far. All of our apps are still in the App Store. Sometimes we disagree, but we’re usually able to talk our way into following what we think are the rules and it seems reasonable.
But sometimes it delays features weeks or months, to be honest. It slowed Tumblr down particularly the last three or four months. We’re trying to launch a lot of user features like tipping and Post+, which are subscriptions. We’re trying to get some of this stuff in there and we’re trying to follow what we think the rules are. If the tips go 100% to the users, we don’t need to charge an App Store fee, which is in many other apps.
I actually did not know that — if you don’t take a cut, Apple won’t take a cut.
That’s the idea, but that feature has been stuck a little bit, so we’re trying to reconcile the UI for that and how it’s presented and everything. It’s just the reality. You live by the App Store, you die by the App Store.
Back in that interview from a few years ago, you were like, “Look, I know people say just make it a web app and screw Apple, but apps are it.” The first half of this conversation, we talked about the glory of the open web and making the open web weird again, making these open-source competitors to these proprietary solution vendors. Are you of two minds? Do you walk around your house just arguing with yourself that you run an app-based business at the mercy of Apple and you also run a completely open-source publishing business over there?
Yeah. I’m a pragmatist. Everything we build is going towards this future of an open web, but also you have to work with the reality of how things are today. Guess what? Every human in the world over the next 20 years is going to have something like smartphones. Those smartphones will have something like app stores and you need to work within that world. I do believe over that timeframe, we’re going to have open alternatives. We need to figure out how to make security and everything else work while we do that. We’ve done it on the web pretty well. We’ve done it for native apps pretty well for things like Mac OS and Windows, so I think it’s possible.
My hope is that the Googles and the Apples of the world — as two of the most valuable companies in the world with cash reserves larger than most countries — will stop acting like underdogs and start acting like the stewards with responsibility to their communities and humanity that they are, and behaving more open.
We’ll always push for that. We’ll always advocate for that. I’m not personally someone who tries to affect change through legislation or lobbying. I prefer to do it in the market.
You’re on the air here a week or so after we had the CEO of Sonos on Decoder. Sonos has obviously gone to Congress. You’re not trying to do that stuff.
Just not my personal specialty.
I think it would be great if somebody would hire a lobbying firm for the open web, so that’s an idea I have for you.
We do our best to represent these issues for the open web. We do try to be present because WordPress is kind of like the dark matter of the web. If you look at the top 10 websites in the world, you’re not going to see the tens and hundreds of millions of WordPress sites aggregated into one name. People forget it exists, but if you added all those up, it might be the largest thing out there. That’s why I call it the dark matter of the web.
We do our best to advocate for that, but we’re also imperfect in that. Automattic and I do not represent the interest of every single independent website and developer out there. I do my best to just say what would make things more open and advocate for that, but I’m not trying to influence competitors or other companies we work with through legislation. That just feels — just not our thing.
Let’s stay on Tumblr for a minute. Tumblr is a social network, and is obviously very consumer. It is very horny. Just being honest about it.
This might be, again, your personal lens.
I don’t even use it. I just hear what people tell me, man. You have moderators, right? How do you moderate the content on Tumblr?
The name of the person I was trying to think of earlier on Tumblr is Amitabh Bachchan.
Amitabh Bachchan. He’s one the most famous actors in India. He’s the Indian actor.
Look at this. He just posted today at 11:42AM. Day 5,130 of his daily updates. It’s awesome. He’s a daily active user of Tumblr.
This is going to turn my mom into a daily active user of Tumblr.
Same thing. One of my best friends Om Malik — you know him. He was one of the first users of WordPress, and has been my best friend forever. He was like, “Matt, this guy is on Tumblr. You’ve got to make a bigger deal of this.”
I’m sure half of the listeners here are losing their minds. He’s a huge deal. How are you moderating Amitabh Bachchan?
You just let him ride.
He moderates you.
The other thing that astounded me about Tumblr. One, that it had some users like this guy who are literally using it every day. Taylor Swift’s there. Most artists’ social networks are managed by a team, but artists are actually using this themselves because Tumblr provides something that’s different from what you’re going to get on Twitter, Facebook, or Instagram.
Number two, 60% of Tumblr users are Gen Z. They are 13 to 24. I had assumed its user base was people who were nostalgic — people who used it in the early or late 2000s or early 2010s were still using it. Some of them were being born then. We have users younger than Tumblr, which is wild.
The other thing that blew me away is that 85% of the usage comes through the app. It’s still getting tens of thousands of signups every day, about half Gen Z, so Tumblr is young. It’s weird. It’s primarily used on mobile, but we also have this power user segment on the web. The web’s still pretty important for us.
Moderating large social networks is honestly really hard. You’ll never hear me piling on to Facebook or other people when they make mistakes moderating, because it’s hard. You have tens or hundreds of millions of people — billions, in Facebook’s case. Everything that happens for humanity happens on your network, and you have to try to make that not go wrong. One of the teams we’ve expanded the most since acquiring Tumblr is support and trust and safety. When we acquired it, Tumblr had a backload of 80,000 support tickets.
Oh my God.
We got that to zero just a few weeks ago.
This is an area that Automattic has a lot of experience with. Our other products are generally well-regarded for not destroying democracy and keeping healthy places on the web. Honestly, the past two years we’ve been doing a lot of catch-up there with Tumblr, and the problem was bigger than I imagined.
It used to be every post we did on Tumblr, people would say, “Oh, you launched this new feature. Why haven’t you gone rid of the porn bots and Nazis?” So we had to do that. There were porn bots and bad people publishing on Tumblr, and we’ve done our best and still today are doing our best, to keep it a healthy, positive place on the web. If I have to say what I would love for Tumblr to be — besides just an alternative, another place you can go that’s different from the other social networks — is a place for art and artists.
Art is necessary for society. It feeds the soul. It’s naturally transgressive. Art pushes boundaries. We need to evolve how Tumblr moderation works to encompass that. It needs to be the best place on the web for art and artists — a place where they can have a direct relationship to their audience and people can follow things, not an algorithm that’s trying to enrage you.
We want people to see things they want to see and get inspired and produce cool things. That’s what Tumblr’s going to be. It’s what it is today for millions of people. We just need to make it much better for that. This is me saying that Tumblr’s moderation policies are going to evolve in a very significant way that I can’t announce entirely yet, but it’s going to get a lot better.
Come on. You were so close. Is it because I called Tumblr horny? Is that what made you change your mind?
At the same time, Tumblr can’t be a place for porn. That’s not what our company can do. It doesn’t work. That needs to be companies like MindGeek or PornHub, and other things that are specifically set up for that. A mainstream company in 2022 with the rules around payment processing, verifying identities, everything. Adult content is rightly having a lot of regulation around it and probably dedicated companies should service that. I can’t say we’re bringing that back. I apologize. I know people want to say that.
It’s the only question on the list, actually. I’m on a long build-up to, “Are you going to bring the porn back?”
But art and artists are really, really strong on Tumblr. I want to make it a really great place for art and artists to thrive and create. That looks a bit more open than what Tumblr allows today, but it’s not the stuff that you would see on a porn site.
Let’s get tactical on that for a minute. Are you going to have a definition of porn? It sounds like you’re going to change the content rules for Tumblr — is that correct? Is it that you’re just not going to announce what they are?
If you look at our other products like WordPress.com, we have policies there that allow a lot more than what’s currently allowed on Tumblr. That’s what we’re going to try to normalize, because those policies have evolved and iterated and worked really well to allow a statue of David or “The Birth of Venus.” Right now that could get taken down — or in old Tumblr it could have gotten taken down. That’s obviously art.
Is that compatible with Apple?
Yes, it’s 100% compatible with Apple.
I think we can do this in a way that also puts a lot more control in the hands of the community to self-moderate and self-tag what they’re publishing in a way that gets the content to people who want to see it, but also protects it from people who shouldn’t be seeing it — people who are under 18, things like that.
We all agree that the app and services should not allow that, but how we implement that I think can be a lot more community-driven. Right now, we have have algorithms that look at things, which is never going to work because, honestly, as a judge at the Supreme Court famously said on the definition of pornography, “I know it when I see it.”
He regretted that, by the way. I just want to point that out.
That judge regretted saying that and he thought it was a mistake.
“It was a real bugaboo for me.”
It is true that writing a strict definition is tricky. I think even when you talk to Apple moderators, they’ll point you to Webster’s definition of it or something.
Yeah, that’s fair.
Some of it is in the eye of the beholder, but there are certain objective definitions that you could say on a piece of content that we could allow better tagging of and then allow filtering of those tags.
Just tactically though, how many moderators do you have working at Tumblr?
There are external moderators as well right now at Tumblr. The actual number is, I believe, above 400.
Okay, and that’s the whole site?
Now, what’s interesting is we actually have a lot less of that on our other services that I think are more effective. Part of what is necessary to create a really effective moderation scheme is building really, really good tools for the moderators to moderate lots of content quickly and to do so in context. Both of those are really, really important for accurate, fast moderation. A single person working an eight-hour day can actually get through thousands or tens of thousands of things if they have the right tools.
I would actually like to decrease our use of external moderation contractors and go to internal teams with good tools to be leveraged here, including using things like machine learning and AI, but not using it for decisions — using it to augment the humans. The center approach that you have probably heard of. AI on its own can be bad, humans on their own aren’t as good as AI at some things, but when you combine them, you can actually get superhuman results — better than either on their own.
One thing we’ve heard about a lot is how bad the jobs are for the people who work in these moderation facilities. They leave with PTSD. Are you thinking about that? Is that something that’s on your mind that you have to solve? Would these tools potentially solve it?
I honestly think about that all the time. In fact, one of the first things I did when coming in to the Tumblr team was do support, and trust and safety, and ride along with folks dealing with that content. I think folks working on this are the unsung heroes of the web as well, because they are not unlike the police force or the veterans that are doing something really tough and hard to protect the rest of us and allow the rest of us to essentially live in a free society. They’re the online equivalents of that, so that’s part of the reason we try to have that team part of the rest of Automattic. I do think it is a challenging job so I have a ton of empathy for it, and that’s why I try to do it myself sometimes too — to both keep that entity strong and also see where we can use software to make it more effective.
I think that’s great. I think more executives at social companies should spend time doing the moderating function. Tumblr is great — the celebrities are using it, the kids are using it. Ultimately, you are in the content moderation business. That’s the heart of the business for Tumblr, for Facebook, for Twitter, whatever, is moderating and ranking content in some way. That’s surprising that you’re doing it, but I’m glad you’re doing it.
Y’all do it too. Your comment section — you have a responsibility to moderate that and keep it a good place you want to be on the web.
We do. Do you know Coral? We have permanent moderators that work with us at The Verge. They’re great. Shout out to Eric. We bought a thing called Coral. If you just look at how the software is expressed, it is a better tool for moderating comments.
It has some user features. It’s meant for that, but the thesis of Coral is that the moderation screen is much easier to use.
Ah. I’ll have to check that out. I will paint a picture: Imagine if The Verge was built on WordPress and Coral was an open-source plug-in. Everyone in the world could get these cool features and the whole web would have better conversations.
The Coral Project guys are going to be all over me when this gets published. That will be great.
I would love for you to try Tumblr again, and send me your feedback. Whatever you want to follow is cool. Whatever bugs you find —
I will put it back on my phone today. My blog was a Tumblog for a long time.
We could spin that back out — a place to aggregate your articles or your podcasts. Just be weird. Be weird on the internet again. It’s fun.
I’ll see if I can open my old Tumblog.
If you post some cool stuff there and send it to me, I’ll boost it as well.
That’s a quid pro quo now. I’ll see. Gotta check the ethics lawyers. I’ve heard from my Gen Z nieces and nephews — they’re all on Tumblr. All their favorite musicians are there.
That’s great. If we can create a third place on the internet that doesn’t have an advertising model — you might have seen that we just launched an ad-free upgrade for Tumblr. Twitter and Facebook never do that because their business models don’t allow them to. But, luckily, since Tumblr isn’t making very much money right now, we can afford to do that and make it the model. I think that’s pretty cool. We have a really decent chance to bootstrap a non-surveillance-capitalism-based social network, which I think is impossible for the incumbents right now. They just have the golden handcuffs.
I wanted to ask about Pocket Casts but I’m running out of time so I will ask you about something else. You’ve been around. You and the web have grown up together. You wrote WordPress. You’re in it. We are in the midst of what some people call a shift to Web3. I feel like you might have many thoughts about the idea that there’s another version of the web coming along. Are you in it? Do you own any blockchain securities? Do you have laser eyes? What’s the situation?
I’m always playing with new technology because that’s just what I do for fun. I’ve tried pretty much everything, including every CMS we’ve mentioned — everything. Every 10 years or so people try to brand a new version of the web. It’s mostly marketing. It’s mostly fundraising, but there’s always some truth to it. Otherwise, it wouldn’t be something we talk about.
It’s also just fun to talk about and debate. “Oh, what is Web3? This isn’t Web3. This is Web2.” I had a thread where I was like, “Everything you’re saying was Web2 actually wasn’t Web2.” It was what came after Web2. Web2 was actually totally open and it had Technorati and Flickr and Delicious. It was all interoperable and had tagging and open content. Then these other things replaced it, like closed social networks.
Web2 was the API era. That’s what I meant. You would have a map and you’d get an API from a thing and then you would mash the two maps together.
Yeah. What’s cool to me is the thing we’re calling Web3 Is actually exactly what we call Web2, but it’s using similar technologies. And I think that’s cool. You know what? Blockchain’s not right for everything, but it’s right for some things.
Name one thing it’s right for, other than currency speculation?
Well, it solves the Two Generals’ Problem. It’s an immutable global record that might be the most secure software that humanity’s ever created.
Okay. I mean, this space is rife with scams, so I buy you in theory, but in practice maybe that’s not the case.
I’m talking about Bitcoin and Ethereum here, what I think of as the top-end blockchains. We’ve never created software this secure for this long.
Microsoft, Apple — billions and billions of dollars put into their operating systems. They regularly have zero days and root exploits. There’s a huge bounty if you’re able to hack Bitcoin and yet it’s still trucking along. I’m astounded at that. That’s incredible.
Bitcoin is also open-source. Another thing, everything I’ve been saying about open-source — so is Ethereum. These are open-source projects. It’s an example of open-source being applied to a new era, which is finance, essentially, and completely transforming it. Open-source is doing something humanity’s never been able to do before. It’s changing the course of things. It’s influencing not just itself, but everything else in the market, which I think is pretty cool.
Now, do I have laser eyes? No. There’s so much to fix with this technology, but I do believe that the problems that people bring up like resource utilization or whatever it is — yes, those are problems, but we’re going to fix them. When I say we, I mean broadly technologists and engineers working tirelessly to release new versions of these things every day, every week, every month, every year.
Do you own any Bitcoin? Do you own any NFTs?
I have a little bit of everything. Actually, WordPress.com was the first major internet service to accept Bitcoin. In 2012, we were on the cover of Bitcoin Magazine.
The article was written by Vitalik, who at the time was the editor of Bitcoin Magazine.
Wow. Vitalik is the inventor of Ethereum, in case people didn’t know. That’s crazy.
I’ve been following this forever. Technologies go through adoption cycles. They go through peaks and trials. They go through winters where everyone thinks it’s the dumbest thing in the world, and that’s actually where a lot of the innovation comes from.
I think they go through high periods and definitely a lot of what’s going on, particularly with NFTs to me right now, feels a little bit like ICOs in 2017 or something like that. You said it, there are a lot of people who are in it for the wrong reasons who are scamming folks. There’s a lot of issues to work out. But, fundamentally, things that give people more freedom, more control, and enable new features for humanity, I would bet on in the long term.
I would just make this point — and I’ll ask it one more time, but maybe with this proviso, I’m asking about Web3. Not Bitcoin, not Ethereum the currency, but Web3. I’d hazily define it as the version of the web where everything you do is a tradable commodity and people are getting paid for all kinds of things vastly more often than now.
That’s what everyone’s getting at. You’re a musician, you release a song. It’s an NFT, people can sell it. When they sell it again, you get a cut. My website has a token and you buy the token. That piece of it where you directly connect the technology to the cultural aspect of the web is what I would call Web3 — if such a definition is even possible to exist.
Okay. So let’s just take that definition as true and I’ll respond to that definition.
Okay. But just to the listeners who are freaking out in their cars, yeah, I just made that one up, but it’s what I got.
That’s part of what makes these things difficult to discuss. People will argue about the definition for hours and hours. But let’s just take the two things you said. What was the first part? Everything is a tradable commodity?
Yeah, everything becomes for sale.
That’s obviously not going to happen. One, I don’t think we would want that. There’s lots of reasons for things not to be tradable commodities. I would say that maybe NFTs will exist, of course, but will everything go that way? Of course not. Two, I don’t think creators getting paid is dependent on blockchain technology. That’s a separate trend that’s happening that is part of the disintermediation of distributors.
Essentially, what the social networks did was they allowed people to go more direct than before. That’s also what the open web’s been doing for 20 or 30 years now. The power is in the hands of creators. I love that salaries are going up for writers and editors and all these sorts of things. This is value going to where it was being created, but that’s happening on WooCommerce, Substack, and other things just as much as it’s happening on any of these “Web3 platforms.”
I think you have to really break out these trends, decide what’s orthogonal and what’s absolutely going to happen. Creators capturing more of the value that they create is 100% going to happen. What’s absolutely not going to happen is everything becoming a tradable commodity.
The financialization of the web is underway, I would say, and maybe it will stop and maybe it will reverse, but right now, all the energy is to financialize a bunch of things.
Just as a counter to that, I will add that one of the most amazing things about the technological revolution was allowing for economics of abundance, not scarcity.
Things get more valuable the more copies there are. We were talking about the positive flywheel of open-source earlier. WordPress gets more valuable the more free copies there are. Now we’re getting more things to introduce scarcity and the value of scarcity into the web, perhaps even programmatically with stuff like NFTs. The difference between what’s come before — from tens of thousands of humanity’s advances — is this idea that, in the world of bits instead of atoms, you and I don’t have a zero-sum way of prospering. We can both benefit from the same thing. We can perfectly copy that software and that actually enables entirely new business models that are pretty exciting. Or maybe that it’s not a business at all, which is okay. Everything doesn’t have to be for profit.
If we have one message to send I think that’s the one we should end with. Matt, you’ve given me some extra time here. I really appreciate it. Thank you so much for coming on Decoder.
Appreciate it. See you in a few years, and hopefully we can grab a drink before then.
Please. I’m dying to.