How to Topple Twitter
Some thoughts on bundling and unbundling social media.
Twitter was supposed to be dead by now.
The servers were going to crash.
The Nazis were going to make the site unusable.
Elon’s “main character energy” was going to alienate the core users.
His selective interpretations of free speech were going to propel anyone with principles out the door.
And yet…
We’re all still there. Why is that?
The short answer is because Twitter does a very specific job for its users and it does that job better than anyone else. If you want to really topple Twitter once and for all, you need to understand what that job is. It’s not enough to appeal to ideology or principle. You have to win on the merits.
And so today, we’re going to look at the world’s online townsquare and how we might be use open technology to replace it. DAOs, open protocols and blockchains all have a role to play. But this isn’t a crypto story. It’s a story about the basics of business — bundling and unbundling.
First, though, we have to answer a question:
What business is Twitter actually in?
Free speech advocates are fond of calling Twitter a global “marketplace of ideas.”
And, indeed, that is the business Twitter is in. But what exactly is a marketplace of ideas?
According to the US Supreme Court, it’s part-and-parcel with America’s commitment to the First Amendment. If we let everyone speak freely, the best ideas will win out. Hence open debate is like natural selection for memes. Only the truest survive.
There’s just one problem. That’s not what a marketplace does.
A marketplace does not decide on the best or “truest” product (whatever that means).
It decides in favor of products that people want. It inspires suppliers to make the things that people want.
A seller in the marketplace of content is someone who creates content in exchange for status. A buyer in the marketplace of content is someone who is looking to consume content that meets their momentary needs. They “pay” the sellers with attention and engagement.
If people want outrage porn, the market will furnish outrage porn until that need is satiated. If people want Nazi propaganda, the market provides, as well. An idea doesn’t need to be true. It just needs an audience’s demand.
Twitter’s value as a marketplace is that it is really, really good at connecting idea buyers (consumers) and idea sellers (creators). It connects them through ranking, through trending topics, through hashtags and through search. Wherever I look, there will be content to consume. Whatever I create, there will be a way to discover it.
That does not enable some high-minded pursuit of truth. It enables the efficient exchange of content between people who want the status that comes from producing it, and people who want the emotional pay-off that comes from consuming it.
Why is Twitter so good at this? Because they have built a robust, vertically integrated bundle that connects creators and consumers.
The Twitter Bundle
Jim Barksdale, the former CEO of Netscape, famously said that there are two ways to make money in business. You can bundle or you can unbundle.
Twitter is a bundler of four main elements.
It has a set of performant, real-time web technology that lets anyone send or consume updates from anyone, anywhere in the world. That’s the real time infrastructure component.
It has a distinctive media format that allows individuals to communicate in short text updates of under 280 characters or an image or a video. That’s the format component.
It has a network/graph of creators and followers who interact with one another. That’s the graph component.
And, finally, it has a set of explicit incentives (ranking, feedback, moderation) and implicit incentives (cultural norms, status-seeking-behavior) that cultivate a culture on the platform. That’s the cultural infrastructure component.
Put together, these four elements Infrastructure, Format, Network and Culture, make the Twitter bundle.
Most of the folks that want to compete against Twitter are trying to offer an alternative bundle. They offer a new format (longer posts! Vertical video!).
They offer a different network (only liberals! Only conservatives! Only Crypto bros!).
They offer a different culture (say whatever heinous racist shit you want! Say only things that are acceptable in a safe space!).
The problem with these bundles is that they assume that they can hold everything else constant while improving one element. But Twitter’s offerings are tightly coupled. You can’t actually re-create Twitter modulo one little improvement.
So you either need to go big – and offer a totally different type of bundle like TikTok – or you need a totally different approach.
The best way to fight a bundler is not with another bundle. It’s to unbundle them.
How do we unbundle Twitter?
Remember: Twitter is a bundle of four things – infrastructure, format, network and culture.
Let’s keep things simple for the moment, and imagine that we want to keep Twitter’s short-text format.
So now, we have three items to unbundle: infrastructure, network and culture. We’re not going to build one company that can compete. We are going to build three new businesses, one for each component of the Twitter Bundle.
The Microblog Infrastructure as a Service
As identified by Farcaster’s Varun Srinivasan, the key features of a common, open social networking infrastructure are the ability to claim a username, to post content and to read content.
That’s it. So our goal is to provide the world’s best infrastructure-as-a-service for microblogging. That means providing for unique usernames and the ability to post/read all the microblogging content that you would want.
Now this is not a new idea. There are a few major players using different technical approaches. We might use a blockchain for this (like Farcaster and Lens intend to do). A blockchain provides an open, decentralized datastore.
Or we could say that’s too expensive, and we could use a federated approach instead. This is what Mastodon does. It allows anyone to run a server that stores data and specifies ways that these servers need to interoperate.
But the problem with both of these infrastructure approaches is… they want to replace Twitter by themselves.
That’s not our game as an unbundler. We just want to provide one part of the stack perfectly.
Our game is to provide a phenomenal microblogging API service that enables you to post to/read any microblogging content in the world. Mastodon and Farcaster don’t have all the content that you want.
Instead they are starting by trying to compete with Twitter. They build apps that encourage people to switch. So you end up with a (very good) Twitter clone like Farcaster or a bunch of Mastodon servers. And those bundles can’t compete with Twitter.
It’s a little like this XKCD comic:
And that’s why the key ingredient that’s missing from these approaches is Twitter itself.
As it turns out, Twitter’s API already allows people to read from and write to Twitter without using its blue-bird app.
So why don’t we bring that in from day 0?
Anyone who publishes to our database can have their content cross-posted to Twitter. Anyone who wants to view content on our app can view from Twitter, as well. Remember – our goal is not to replace Twitter, it’s to unbundle its infrastructure.
Our goal is to build an infrastructure that captures, stores and accesses all of the microblog posts in the world. So we are going to pull in content from Twitter, Farcaster, Mastodon etc. We are going to make that aggregated content open and universally accessible.
This approach is not particularly novel – Plaid does some version of this with its banking plug-ins, Matrix and Texts.com do this with messaging. It also happens to be how Instagram grew by allowing users to publish into Facebook.
The idea is simple: create a single, globally available access point for all microblogging content while building your own reservoir of open content. Make that content available to anyone who wants to access it. Support yourself via some kind of protocol revenue share, API fee, etc. That’s the service.
Network-as-a-Service
In my early days at Facebook, there was a sign plastered on the wall that read: “People don’t use Facebook because they like Facebook, they use Facebook because they like their friends.” It’s an elegant reminder that the ultimate value of a social network is the people who use it.
Facebook once imagined that its social graph would form the backbone of a new internet. But the shift to the iPhone changed that. The rolodex in your life stopped being your Facebook friends and started being your phone’s contact book. This is why, today, when you login to a new mobile app it asks you to connect your phone book rather than your Facebook friend list.
For parasocial follower relationships, the calculus has also started to shift. Substack constantly reminds creators that its value is in “owning your audience.” When you own their email address, you have an unfiltered channel to access your users. Twitter and Facebook can’t take your audience away.
Our eventual goal is to build a network-independent graph. This means that a user can login to any app, find anyone they want to follow, and receive updates from them there – much like you can choose to receive emails via Gmail or Outlook. It’s hard (and a violation of privacy!) to reach people on a platform they didn’t select.
So what if, instead, we start from the consumer side. We let a user import their contacts, their email address book, their Twitter following list or Facebook friend list. And we let them manage that directly from their CRM – we let them choose where and how to receive updates from the people in their contacts. We let them know where else that contact is creating content that they can access. We let them pull it into their client just like it’s email.
We make it easier for that user to decouple their contacts and their contacts’ content from the specific platform that they are connected to.
And we tell creators that if they want to assure reach to fans, they should encourage their fans to follow them via our platform. It will ensure their fans get updates on Twitter, Instagram, YouTube, Email or anywhere else.
We become the one surefire way to assure reach between different users of many social networks. That won’t get us all the way there. But it’s a start to a platform-independent-graph.
Cultural-Infrastructure-as-a-Service
Today, culture is tightly coupled to platform. There are posts that make sense on Twitter that would not make sense on Instagram and vice versa. But this means that folks who want to consume all the content from their creator need to buy into the cultural norms of that platform. And maybe I really like memes on Twitter, but I don’t want to support Elon Musk’s political agenda.
When I read an email or a text message, I don’t have to worry about what that says about me. A neutral medium gets out of the way. A good protocol should do that. Its job is to carry information between individuals. But – the thing is – there’s a lot of content out there.
And Twitter’s key value is in connecting the supply of content with the demand of consumers. Ranking, moderation, UI design, search/sort features – all of these help determine what content gets surfaced and thus influence the culture of a space. They are the cultural infrastructure of social media.
Although Twitter is a marketplace for ideas, it is a centrally planned culture. There is only one feed ranking. There is only one trending algorithm. There is only one search algorithm. There is only one content-moderation policy (though it changes depending on the whims of the CEO). We can change that with our unbundled approach.
Open data and open graph means that you can choose the cultural infrastructure that you want. You can choose a content moderation service with specific policies. You can choose a ranking algorithm (or algorithms!) that meet your needs. Some days you might want an AI trained to inform you, sometimes you might want one that shows you the cutest puppies.
Rather than a single marketplace of ideas with a single matching-mechanism, we can let a thousand marketplaces bloom. And that’s the secret to this whole strategy.
Twitter is a marketplace for content. It connects people to the text/link content they want efficiently. To beat it and topple it, you will need more than an ideological appeal.
You need a system with more content, more reach and better matching mechanisms.
The best way to achieve that is not to build a better Twitter. It’s to build an ecosystem of services that render Twitter as obsolete as the Internet rendered the siloed dial-up services of old. Then, when you post about how Twitter is dying, you might finally do it on a service other than, y’know, Twitter.