You don’t have to be an engineer to feel the difference. Open a messaging app, and you’re either in a walled garden or on a public footpath. The garden is pretty, well-maintained, and has a gatekeeper. The footpath is a little rougher, but nobody can fence it off. That’s the gap between a platform and a protocol, and it’s not just a technical detail. It’s a decision about who gets to make the rules. Maren Solberg has spent twenty years watching these decisions get baked into the systems we live inside, and the pattern never changes: platforms hoard power, protocols spread it around.
What a Protocol Actually Is
A protocol is a handshake. It’s a shared set of rules that lets different systems talk to each other without anyone asking for permission. Email runs on SMTP, a protocol hammered out in 1982. The web runs on HTTP. These aren’t products you buy from a company. They’re specifications—documents that say, “If you format your data this way, everyone else who follows the spec will understand you.” You can build software that speaks SMTP, and it’ll exchange messages with any other SMTP-speaking software on the planet. Nobody owns the protocol. The Internet Engineering Task Force keeps the documents tidy, but the protocol itself belongs to whoever uses it.
That’s why your Fastmail account can chat with a Gmail account without a second thought. The two companies compete, but they interoperate because they both follow the same rules. A protocol doesn’t care about your business model. It doesn’t care about your market share. It just defines the grammar of the conversation and gets out of the way.
What a Platform Is
A platform is a fenced-in field. It’s a piece of software—or an entire ecosystem—owned and operated by a single entity. Facebook is a platform. Twitter is a platform. Uber is a platform. When you step onto a platform, you’re not just using a tool. You’re entering a jurisdiction where the owner writes the laws, holds the keys to your data, and can redraw the boundaries whenever they like. Platforms often start by piggybacking on open protocols. Facebook leaned on HTTP and email notifications to pull people in. But once they reach a certain size, they brick up the exits. They stop playing nice with competitors. They make leaving feel like a punishment.
The real question is who holds the power. In a protocol-based system, power is scattered across everyone who participates. Run a server, build a client, spin up a new service—nobody can stop you. In a platform, power pools at the top. The owner decides what’s allowed, who gets silenced, and how your information gets used. If you don’t like it, your only move is to walk away—and lose everyone who stays behind.
The Email Story: A Protocol That Refused to Die
Email is the old warhorse that keeps proving the point. SMTP was born in 1982, and it still carries the world’s messages. You can pick from thousands of providers. You can run your own server in a closet. If you own your domain, you can switch providers without changing your address. The protocol doesn’t care who you are or what client you use. It just moves the mail.
But email also shows how platforms try to colonize open ground. Gmail is the biggest player, and Google has spent years grafting on proprietary features that only work inside Gmail. Smart replies, nudges, certain spam-filtering tricks—none of that is in the SMTP spec. Those are platform features, designed to make Gmail stickier. The protocol still hums along underneath, but the platform layers on top create a quiet lock-in. Most people don’t notice until they try to pack up and leave.
Social Media: The Platform Counter-Revolution
Social media could have grown up on protocols. For a while, it did. RSS let anyone subscribe to updates from any blog or news site. XMPP and IRC ran open, decentralized chat. But these protocols didn’t have venture capital pouring gasoline on them. They didn’t have growth hackers or engagement algorithms. They were slow, steady, and boring—exactly what good infrastructure should be.
Then Facebook, Twitter, and Instagram arrived. They traded convenience and network effects for control. They made it dead simple to find friends and share photos, but they also made it impossible to take your social graph with you. Your followers, your posts, your identity—all locked in a company’s database. The platform model didn’t win because it was better technology. It won because it was politically convenient. It concentrated power in a way that made rapid growth and monetization easy.
Why Protocols Are a Slog to Build
Protocols demand coordination. You need a bunch of independent parties to agree on a standard, implement it, and keep things compatible. This is slow, messy work. It means committees, mailing lists, and endless arguments about weird edge cases. Platforms, by contrast, can sprint. A single company can design, build, and ship a feature in weeks. They don’t need consensus. They just need users.
But that speed has a dark side. When a platform screws up, everyone on it pays the price. Facebook tweaks its algorithm, and millions of small businesses watch their reach evaporate overnight. Twitter suspends an account, and there’s no real appeal. The platform’s word is final because there’s no other court to go to. Protocols, by spreading power around, also spread the risk. If one email server goes down, the rest of the network keeps humming. If one XMPP server starts censoring, you can move to another and keep your contacts.
The Political Layer Beneath the Code
Every piece of infrastructure has a political philosophy baked into its bones. Protocols are federalist. They set up shared rules with local autonomy. Platforms are authoritarian. They centralize authority and expect obedience. This isn’t a metaphor. It’s a structural fact. When you build on a protocol, you’re choosing a governance model. When you build on a platform, you’re choosing a boss.
Maren Solberg has watched this same story play out in city planning, energy grids, and transportation networks. The dynamics are identical. A public road is a protocol: anyone can use it, and no single entity controls access. A private toll road is a platform: you pay for entry, and the owner sets the terms. The technical layer is just the surface. The real question is always about power.
When Platforms Swallow Protocols
There’s a pattern that repeats across the tech industry like a bad song. A new technology shows up as an open protocol. Early adopters build tools and services on it. Then a well-funded company steps in, builds a slicker user experience on top, and starts adding proprietary extensions. Over time, the protocol becomes a commodity, and the platform becomes the tollbooth.
This happened to the web itself. HTTP is an open protocol, but Google’s search engine became the de facto index of the web. Chrome became the dominant browser. AMP, a Google-controlled framework, tried to replace HTML for mobile pages. Each move pulled more of the web’s infrastructure under Google’s roof. The protocol still exists, but the platform has captured so much traffic and tooling that the openness is increasingly theoretical.
The Exit Strategy Problem
Here’s a quick test to tell a platform from a protocol: Can you leave without losing everything? With email, the answer is yes—if you own your domain. With social media, the answer is almost always no. Your data, your connections, your reputation are all bolted to the platform. This is intentional. Platforms spend heavily to make exit painful. They call it “retention,” but it’s really a form of lock-in.
Protocols, on the other hand, have low exit costs. You can switch email providers, DNS registrars, or web hosts without breaking the underlying system. The protocol guarantees continuity. That’s why protocols are tougher. They don’t depend on any single company staying alive. When a platform dies—remember MySpace, Google+, or Vine?—everything built on it vanishes. When a protocol-based service shuts down, the network keeps going.
Building for the Long Haul
If you’re an engineer or a founder, the choice between building on a platform or a protocol is one of the biggest decisions you’ll make. Building on a platform is faster and often more profitable in the short run. You get access to existing users, payment systems, and distribution channels. But you’re also building on rented land. The platform can change the rules, jack up prices, or compete with you directly. Amazon has done this to countless third-party sellers. Apple has done it to app developers.
Building on a protocol is slower and harder, but it creates value that lasts. You’re contributing to a commons that can outlive any single company. The Linux kernel, the HTTP protocol, and the Bitcoin network are all examples of protocol-based infrastructure that has survived for decades. They’ve weathered corporate attacks, government pressure, and internal fights. They persist because no single entity can pull the plug.
FAQ
Is Bitcoin a platform or a protocol?
Bitcoin is a protocol. It defines a set of rules for a decentralized ledger, and anyone can implement those rules in software. There are multiple Bitcoin clients, multiple mining pools, and multiple exchanges. No single entity controls the network. However, some services built on top of Bitcoin—like certain exchanges or custodial wallets—are platforms. They add a layer of control on top of the open protocol.
Can a platform become a protocol?
Rarely, and usually only under pressure. When a platform becomes too dominant, regulators or market forces may force it to open up. Microsoft’s move to support open document formats was partly a response to antitrust pressure. But true protocols are designed to be open from the start. Retrofitting openness onto a platform is difficult because the business model depends on control.
Why do users choose platforms over protocols?
Convenience and network effects. Platforms invest heavily in user experience, making their products easy to use and full of features. Protocols often have clunkier interfaces because they’re built by volunteers or small teams. Platforms also benefit from the fact that everyone else is already there. It’s hard to switch to a protocol-based alternative when your friends, family, and customers are all on the platform.
Are there any successful protocol-based social networks?
The Fediverse—a network of servers running software like Mastodon, Pleroma, and Pixelfed—is a protocol-based social network built on the ActivityPub standard. It’s small compared to Facebook or Twitter, but it’s growing. The key difference is that no single company owns the Fediverse. You can choose a server, move between servers, and still interact with anyone else on the network. It’s social media the way email is email.