In September 2025, Alison Roman announced that she had moved 343,000 newsletter subscribers from Substack to Ghost. The following month, Anne Helen Petersen—one of Substack’s most prominent writers and the recipient of one of her six-figure growth paths—took to Patreon. Lyz Lenz followed suit, arguing that bot subscribers were artificially depressing her engagement rates and that Substack’s algorithm was pushing what she described as “anger, Nazis, transphobia and conspiracies.” For two weeks on Patreon, 70% cured its paid subscriber rate.
These were not isolated cases. According to an industry piece that tracks location, Beehiiv has seen the migration of around 3,000 creators From Substack for twelve months to March 2025. Ghost has been the destination of choice for publishers who want control over their brand and technical infrastructure, while Patreon has attracted writers motivated mainly by content moderation concerns and audience quality issues.
Taken individually, each move has its own rationale. Taken together, they expose something more structural—a tension at the heart of the promise to “own an audience” on which the newsletter platform era was built.
What “having an audience” really meant
Substack’s original sound was a direct response to the broken promise of the social media era. Facebook and Twitter taught publishers an expensive lesson: build on someone else’s platform, build an audience there, and watch the algorithm change conditions until your reach collapses. Substack offered an alternative framework: you own your email list, you own the subscribers, and the platform is just infrastructure.
This framework was and remains partially accurate. Substack allows writers to export their subscriber lists. Migration to Ghost is technically simple; Alison Roman was able to transfer 343,000 subscribers without breaking a single subscription because both platforms process payments through Stripe. An email list is, in the narrow technical sense, portable.
But moving a list isn’t the same as having an audience. And the migrations of 2025 reveal exactly where the gap is.
Writers who can leave Substack and take their audience with them are almost invariably writers who have already built a “follow-you-anywhere” following. They don’t need Substack’s discovery engine. They don’t need the Notes algorithm. Readers subscribe because of who they are, not because Substack exposed them.
For everyone else—the majority of publishers on the platform—the audience relationship is more muddled. Growth on Substack is increasingly driven through Substack Notes, through the platform’s recommendations engine, through algorithmic boosting that rewards certain token types. A writer described Substack’s partner said that the success team needs to send a special message three times a day to satisfy the algorithm. An email list can be portable. It is not the growth mechanism that builds it.
As success grows, so does the problem of payment
The economics of a 10% substack cut are designed to be reasonable at startup and worry at scale – that’s when it matters most.
When a writer has 50 subscribers paying $8 a month, $40 going to Substack is an abstraction. When that writer has 2,000 paying subscribers at $8 a month, the $1,920 monthly fee is the cost of a platform that increasingly competes with its own content for attention with its social features. As noted by one writer who moved to self-hosted Ghost: At 100 subscribers paying $10 per month, Substack is already charging $100 per month – more than flat-fee platforms charge for lists of comparable size.
Ghost charges a fixed monthly fee regardless of income. In meaningful subscription income, the financial situation for migration becomes a simple calculation. Ghost’s own comparison calculator makes the point clear: a publisher with 1,000 subscribers paying $5 a month earns $60,000 in annual revenue, $6,000 of which goes to Substack. Ghost’s fixed fee of this magnitude is part of that.
Brad HargreavesRunning Substack’s top-grossing real estate newsletter, he moved to Ghost in September 2025 primarily for capability reasons, not payment: he needed API access and web hooks to integrate his courses and database products into his publishing business. Substack’s closed platform failed to accommodate the work it was trying to build.
This points to the second dimension of the ownership problem. Substack’s infrastructure is optimized for a specific newsletter publication. Writers who wanted to grow into more complex businesses—courses, databases, multi-level memberships, custom integrations—finally hit the ceiling. As an open source platform, Ghost has no comparable ceiling.
A measure of moderation that platforms will not address
A persistent theme running through 2025’s departures is content moderation—specifically, the concern over sharing a platform and recommendation engine with writers who object to the content of many independent publishers.
This is of importance beyond the ethical dimension. The recommendation engine in Substack connects publications with potential subscribers on the platform. The newsletter may be recommended alongside or against content that the writer deems harmful. Platform Notes features posts from across the ecosystem. For writers whose audience relationships are built around a certain set of values, existing within a platform that monetizes and algorithmically boosts content that goes against those values is not just a philosophical problem, it’s a brand positioning problem.
Nieman Lab’s coverage of departures in October 2025 noted that this wave of exits also showed email delivery failures and lack of technical support as practical complaints. As Nieman points out, the “Why I’m Leaving the Substack” post has become a genre in its own right — and the reasons have evolved from the early days of “10% is too much” to a broader dissatisfaction with what the platform has become at scale.
What the migration pattern actually reveals
The uncomfortable truth that goes into all of this is that “owning an audience” has always been more contingent than in the days of the newsletter platform. The email list is active. But the relationships in this list – trust, engagement rate, regular attention of the subscriber – were built partly through the platform’s infrastructure and partly through the writer’s own work. Separating these contributions is more difficult than exporting a CSV.
The writers who left Substack for Ghost in 2025 were mostly those who had gained enough independent credibility that the platform’s contribution to their audience relations had become marginal. They may migrate because over time they have become less and less interested in why their readers subscribe to the Substack infrastructure. The writers who can’t leave so cleanly are the ones whose growth is still entangled with the platform’s discovery and recommendation systems.
This is a dynamic that every publisher built on any platform eventually faces. The substack is not uniquely to blame. The same analysis applies to any platform that offers distribution in exchange for revenue share or audience data. Terms seem generous early on, when distribution is a scarce resource. Once they build something that needs the platform as much as the publisher needs the platform, they look different.
The lesson for bloggers and independent publishers is more about what platform to choose than what is required to “have an audience.” A portable email list is a starting point, not a destination. A real asset is an audience connection so strong that readers will follow a publisher anywhere – across platforms, through migration, whatever algorithm is in vogue at the moment.
Building is slower and more difficult than optimizing for a search engine. It is also the only version where audience ownership cannot be taken away.
