How to Kill Your Own Platform in Three Easy Steps (Companies Keep Trying)

Imagine you've been handed a golden goose. It's healthy, productive, beloved by everyone who's seen it, and requires relatively modest upkeep. Now imagine that every year, a new group of executives files into the room, stares at the goose for about six months, and then starts discussing whether the goose would be more valuable if it were slightly smaller, angrier, and surrounded by advertisements for things it doesn't want.

This is the story of online communities in America. It has been the story for about fifteen years. It keeps happening. Nobody seems to learn anything.

Let's talk about why.

The Tumblr Playbook, Annotated

Tumblr is the canonical example, which is why we're calling this the Tumblr Principle, but it's worth being precise about what actually happened there — because the lesson is more specific than "don't be Tumblr."

Tumblr in its prime was a genuinely strange and generative place. Fan communities, artists, writers, political activists, LGBTQ+ teenagers finding their people, niche humor that made no sense outside of its own context — it was chaotic, deeply weird, and fiercely loved by the people who lived there. Verizon acquired it. Verizon tried to make it advertiser-friendly. Making it advertiser-friendly meant making it content-moderated in ways that were simultaneously too broad and too inconsistent. The NSFW content ban in 2018 was the famous turning point, but the damage was cultural before it was statistical: the platform signaled, loudly and clearly, that it did not understand or trust its own users. The users left. The platform is now owned by Automattic and is a ghost town with great bones.

The Tumblr Principle, distilled: communities don't leave because they find something better. They leave because the platform makes them feel unwelcome in their own home.

Twitter: A Master Class in Accelerated Self-Destruction

If Tumblr is the slow burn, Twitter — now X, a rebrand that itself communicates contempt for the platform's history — is the case study in what happens when you pour gasoline on the slow burn and then tweet about it.

Elon Musk's acquisition of Twitter in late 2022 and the subsequent eighteen months of policy chaos, mass layoffs, advertiser exodus, verification scheme reversals, and algorithmic changes so frequent they felt random — all of this has been documented exhaustively. But the most instructive part isn't the drama. It's the pattern.

Musk inherited a platform with a deeply entrenched user base, a recognizable cultural role, and genuine network effects that made it difficult to replace. These are enormous structural advantages. He proceeded to make decisions that systematically antagonized the exact users who gave the platform its cultural cachet — journalists, academics, brands, artists, public figures — while simultaneously making the product worse for everyone else through chaotic moderation and a timeline algorithm that increasingly prioritized his own posts.

Advertisers fled. Engagement metrics became unreliable. Users migrated to Bluesky, Mastodon, Threads, and good old-fashioned not-being-online. The platform still exists and still has users, but it has lost the thing that made it Twitter: the sense that being there meant something.

You cannot buy cultural relevance. You can only inherit it, and then spend it.

Reddit and the API Rebellion

Reddit's 2023 API pricing crisis is a slightly different flavor of the same syndrome, and it's worth examining because it almost went differently.

Reddit announced changes to its API pricing structure that would effectively kill third-party apps — apps that millions of users, and crucially, many moderators, depended on to make the platform usable. The backlash was immediate and organized. Hundreds of subreddits went dark. Moderators coordinated. Users made noise.

Reddit's leadership, to the surprise of basically nobody who had been paying attention, initially held firm and then made targeted moves to undermine the protest — replacing moderators, adjusting policies to force subreddits to reopen. The third-party apps died. Apollo, the beloved iOS client, shut down. The promised IPO timeline created pressure to show advertisers a platform that was under control.

But here's the thing: Reddit survived. Not because it handled the situation well — it didn't — but because Reddit's communities are deeply embedded in the platform in ways that Tumblr's weren't. The content is the archive. The search traffic is enormous. Leaving Reddit means leaving years of accumulated conversation that doesn't exist anywhere else. The switching cost is high enough that users stayed even after being treated poorly.

For now.

Why Does This Keep Happening?

The pattern is so consistent that it demands a structural explanation. These are not stupid people running these companies. They have access to the same user research, the same engagement data, the same cultural commentary that everyone else does. They can see what's happening. So why do they keep doing it?

A few reasons, none of them comforting.

First, the people who make these decisions are almost never the people who use the product the way its community does. Executive leadership at a platform company tends to experience the platform as a data dashboard, not as a place. They see monthly active users and ad revenue and content policy compliance rates. They do not see the inside jokes, the years-long friendships, the elaborate community norms that make a place feel like a place. You cannot optimize for something you cannot perceive.

Second, the incentive structures are misaligned in a specific and predictable way. Community health is a long-term asset. Quarterly revenue is a short-term metric. When the two come into conflict — and they always come into conflict eventually — the quarterly number wins, because the quarterly number is what gets measured, reported, and compensated.

Third, and most cynically: once a platform is large enough, its leadership often concludes that the community is captive. Users have history there. They have followers, archives, connections. The switching cost feels prohibitive. So the platform tests how much it can take before users leave — and sometimes it miscalculates, and sometimes it doesn't, and either way the lesson is the same: you can extract a lot from people before they walk away. The problem is that "a lot" has a ceiling, and companies reliably push past it.

The Goose, Again

Every community platform starts as something people chose freely. They showed up because it was good, or interesting, or theirs. The platform's job, in theory, is to protect the conditions that made people want to show up in the first place.

Instead, the job becomes extracting value from the people who showed up — their attention, their data, their content, their loyalty — until the extraction exceeds what the community can absorb. At that point, the goose is gone. The executives who made the calls have usually already moved on to the next opportunity. The users are left with a platform-shaped hole where something they cared about used to be.

This will happen again. It's probably happening somewhere right now. The internet is broken. We took notes. The goose could not be reached for comment.