Substack's 10% and What You're Actually Paying For

The fee is simple. The trade behind it takes longer to explain.

What the 10% Buys

Substack's pricing is straightforward on the surface: the platform takes 10% of every paid subscription, plus Stripe's payment processing fee of around 2.9% and 30 cents per transaction. A writer charging $10 a month keeps roughly $8.40 of it. At low subscriber counts that arithmetic barely registers. At scale — tens of thousands of paying readers — it becomes a meaningful cost centre.

The 10% funds a specific bundle of infrastructure. Substack handles hosting, email deliverability, payment collection, subscription management, and a discovery layer through its own recommendation network. None of those things are trivial. Email deliverability in particular is an ongoing technical operation: maintaining IP reputation, managing bounce rates, handling unsubscribes in compliance with CAN-SPAM and GDPR, and navigating the SPF, DKIM, and DMARC authentication chain that determines whether a message lands in the inbox or the spam folder. A writer sending independently must either contract those services separately or develop the operational knowledge to run them.

Substack also provides an RSS feed for every publication automatically. Paid posts can be paywalled within the feed, but the feed itself exists, which matters to readers who prefer a reader client over email. The platform's architecture treats email and RSS as parallel delivery channels, not competing ones — a posture that distinguishes it from pure email-marketing tools with no feed output at all.

The Distribution Question

The more interesting part of the 10% trade is what it buys in terms of audience mechanics. Substack's recommendation system, which surfaces one writer's publication to another's subscribers at the moment of signup, functions as an internal discovery network. It is a closed system — it works only if you are on Substack — but for newer writers it represents genuine reach that would otherwise need to be purchased through advertising or built through manual outreach.

That discovery value is real but asymmetric. Writers with large existing audiences import their own readers and gain little from the recommendation graph; they are effectively subsidising its operation for smaller publications while seeing limited benefit themselves. For them, the 10% is closer to a pure infrastructure fee. Whether that fee is competitive depends on what the alternative costs.

Moving between readers is a file transfer, not a migration. OPML carries the URLs; read state, starred items and folders mostly stay behind.Photo: SHVETS production / Pexels

The Ghost Comparison

Ghost, run by the Ghost Foundation as a non-profit, offers a materially different model. Ghost publishes a feed by default and treats newsletter delivery as core to its open-source architecture rather than a platform-specific feature. On Ghost's managed hosting (Ghost Pro), pricing is tiered by the number of members rather than a revenue percentage — at lower tiers the monthly fee may be less than Substack's cut once a publication is earning meaningfully. At higher subscriber counts the economics typically favour Ghost.

The difference sharpens when you consider portability. A Substack publication's subscriber list is exportable as a CSV, but the recommendation-network relationships, the SEO equity of a substack.com subdomain, and the payment infrastructure are not. Moving means rebuilding on each of those dimensions. Ghost's self-hosted path, where the writer controls the domain and the database entirely, leaves none of those relationships in a third party's hands — though it requires the writer to manage or pay for deliverability separately through a transactional email provider.

Ghost's model shifts the operational burden onto the writer but preserves the relationships — with readers, with the domain, with the subscriber list — under the writer's own control.

The framing that matters is not "which platform is cheaper" but "what dependency are you acquiring." Substack's 10% purchases convenience and a discovery layer; it also purchases lock-in that compounds as a publication grows. Ghost's model shifts the operational burden onto the writer but preserves the relationships — with readers, with the domain, with the subscriber list — under the writer's own control.

The Honest Calculation

Neither model is wrong. Substack's fee is defensible for writers in early growth phases who need infrastructure without operational overhead. The recommendation network genuinely moves subscriber counts for publications that engage with it. But the same writer at a hundred thousand paid subscribers is paying a sum that funds a full engineering team, for a service that has not grown proportionally more complex to deliver.

The open-web calculus is that every percentage point paid to a platform is a percentage point not spent building infrastructure that travels with you. The question each publisher answers differently is how much that portability is worth before the compounding begins.