From locked‑in to free to leave: The case for digital platform interoperability 

Have you ever wondered why everyone uses the same handful of social networks? You may not love X, Facebook, Instagram, or LinkedIn, but leaving means abandoning your friends, colleagues, or professional networks. So, you stay. This hands power to the platforms: users must accept whatever rules, moderation, and algorithms a provider imposes.  

It doesn’t have to work this way. Consider the telephone: you choose Vodafone, your friend chooses O2, and you can still call each other. Or e–mail: a @gmail.com address can write to a @protonmail.com one without difficulty. Economists and regulators name this principle “interoperability.”  

Imagine it applied to social media: if all your friends are on Instagram, and Instagram were required to interoperate, you could decide to join a competitor (call it “Freepixel”) and still see, follow, and talk to everyone on Instagram.  

This idea of social media interoperability promises to distribute platforms’ power by re-enabling competitive market forces without sacrificing users’ connectivity. It has gained serious momentum: theoretical economics literature, legal analyses, and government expert reports recommend it.   

Accordingly, policymakers have begun to act: the EU’s Digital Markets Act has mandated interoperability for messaging services (e.g., WhatsApp) since 2024 and requires a periodic review of extending this to social networking. Comparable proposals are also advancing in the U.S. (Utah, New York, Federal).  

Industry-funded think tanks, on the other hand, warn that forcing social media interoperability would create “disproportionate technical, security, and societal risks”. And yet, the debate lacks crucial empirical evidence on whether social media interoperability would live up to expectations. 

How do you research a policy that hasn’t been implemented? 

A reason for this evidence gap is that no mainstream digital platform has been required to interoperate yet. Expectations have instead been extrapolated from experiences in older network markets like telephone and email systems. Whether and how those lessons carry over to social media remains an open question. 

My research looked for corners of the internet where social media interoperability already operates, and where its practical effects can be observed. One such space is the Fediverse with its Mastodon software: it enables an interoperable, Twitter-like microblogging network with roughly 740,000 active users. Those users are spread across more than 8,000 independent providers, called “instances”, that interconnect through a shared protocol. Each instance is a self-contained service, much like Vodafone or O2, yet they exchange data with one another.  

Unlike on conventional platforms, where leaving means abandoning everyone, Mastodon lets users switch providers while staying within the same userbase and without losing followers. Mastodon users even post openly about why they are leaving and what they are looking for instead.  

I collected millions of these posts from across thousands of providers, identified the users who switched, and analysed their reasons and experiences.  

Three lessons from Mastodon for social media interoperability 

My research is among the first to provide empirical evidence on user choice under social media interoperability. I find that it can give users a wide range of provider choice, but that it only delivers on its expectations when it accounts for the specific ways social media differs from earlier interoperable markets. 

First, social media interoperability needs to extend beyond interacting with people you already know. In legacy networks, communication was “tie‑based”: you contacted specific people.  

But what makes social media specific is “open‑network” interactions: discovering strangers’ posts, joining wider conversations, or searching across the network. Here, Mastodon’s interoperability is incomplete: not for technical reasons, but because Mastodon is built mostly by volunteer open-source contributors and a tiny paid team. This meant users moved toward larger providers, because on smaller ones, they felt like missing out.  

The lesson for policy and developers is that interoperable social media must support both “tie‑based” and “open‑network” interactions. If interoperability is only partial, there might still be a pull towards larger providers. 

Second, frictions in choosing and switching providers remain when “user assets” and “provider services” are bundled together. On Mastodon, users can move their followers across providers, but not other “user assets”, such as their handle, post history, or community membership. Meanwhile, “provider services”, such as moderation rules, provider jurisdictions, and service levels, come as one fixed package. This creates information costs when users search for the right bundle, and switching costs when moving to another one.  

The lesson is that all “user assets” should be portable, and “provider services” more modular. Bluesky, another interoperable social network, shows that such modularity is technically feasible. 

Third, interoperability actually allows meaningful differentiation between providers. Some critics warn that the need to follow a standardised technical protocol would make providers indistinguishable.  

My evidence shows the opposite: providers vary vertically in quality, and users can hold providers accountable by leaving poor performers. They also vary horizontally, such as in differing rulesets, moderation philosophies, governance models, or hosting jurisdictions. This pluralism lets users choose a provider whose governance matches their preferences, something no centralised platform can offer. 

The specifics make or break the mandate 

Social media interoperability can expand user choice, but social platforms pose challenges that the experience from legacy networks cannot anticipate.   

My research highlights three areas specific to social media interoperability: “tie‑based” vs “open‑network” interactions, “user assets” vs “provider services”, and horizontal vs vertical differentiation. These specifics matter for platform designers and policymakers alike. If they are underestimated, regulators may be underprepared for effective enforcement, well-intentioned mandates may fail to deliver, and large platforms will be quick to claim: the effort was misguided. 

This blog post is based on Riederle, P. (2026). “Does Digital Platform Interoperability Deliver on Its Expectations? Evidence From User Switching on Mastodon.” (Under Review, Preprint available). This research was recognised with the “Jovana Karanovic Impact Award”. 

Find out more about Philipp’s research interests.