The rake and the network

A woman I know left a staff job at a magazine in 2023 to write a paid newsletter, and at dinner last month she showed me the numbers the way people show you a baby. Several hundred paying subscribers, a number she would have been fired for naming inside the magazine, real money, deposited monthly. Then she walked me through her morning. Not the writing morning, the other one. She answers comments in the app, posts a chat thread, checks what Notes did overnight, replies to three recommendations, and looks at the conversion dashboard, and only then, around eleven, does she open the document where the writing happens. She said something I've been turning over since. The writing is maybe twenty percent of the job now, she said, and the job has a boss, and the boss is a feed.

She's a success story. Substack's own materials would put her in a testimonial, and the testimonials are accurate. The scene is the argument anyway: the platform sells independence and runs a network, and the two promises pull against each other in ways the sales copy doesn't mention. Substack is the most consequential publishing company of the last decade and I want to take it seriously, which means taking its incentives seriously.

The rake is the business model

The economics are public and unusually clean. Substack is free to start, free to leave, and takes ten percent of subscription revenue, minus card fees. The company reported five million paid subscriptions in the spring, its own count, and the number is credible against everything else I know. For a writer starting from zero, this is a fair deal: ten percent of revenue buys payments infrastructure, email delivery at scale, hosting, and the absence of any technical work. That's a real service and the alternative is not free.

The fee's structure is the interesting part. It scales with your success and never ends. A writer billing a hundred thousand dollars a year pays the platform ten thousand, every year, for infrastructure that a flat-rate competitor sells for the price of a dinner. The crossover point exists and it's computable, and almost nobody computes it, because computing it requires imagining your own success and then doing arithmetic about a person you aren't yet. In my consulting years I watched a software company price almost exactly this way, a percentage of the client's revenue that looked small at the sales meeting and became the client's largest line item by year four, and the client never once renegotiated, because by then the switching costs had been built for them, cheerfully, one integration at a time.

The network grows you, and then you're grown

Consider the incentive in the growth claims. Substack says more than half of new subscribers now arrive through its built-in network: the app, Notes, recommendations, the leaderboards. It also says over thirty percent of paid subscriptions come from within that network. Both numbers come from the party selling the network, so treat them as an advert with a decimal point in it. But take the shape as true. The platform has become a distribution channel, the thing that decides which new voices get seen, which is the role magazines used to hold and radio before that.

A distribution channel is a policy instrument, whether or not it wants to be. The moderation fight of late 2023, when writers publicly left over the platform's refusal to act on certain newsletters, wasn't a free speech dispute so much as an argument about who the algorithm serves. Substack's laissez-faire posture is itself a distribution strategy; it keeps the audience that others would eject, and the audience is the asset the company has been assembling since it raised its first venture money. The company has since raised round after round, from Andreessen Horowitz and others, into a valuation around a billion dollars as reported in 2025. Venture money wants a return, returns want growth, growth wants engagement, and engagement in this context means time inside the app, reading things the writer didn't choose to send. (The company's history and funding are summarized here.)

What "you own your list" actually buys

The company's best defense is that you can leave, and it's stronger than its critics admit. Substack lets you export your subscriber list and your archive, which means the lock-in is soft by the standards of the industry. A writer who leaves takes the addresses with her. This is genuinely unusual, genuinely valuable, and worth more than the fee in pieces of mind.

But look at what the list doesn't include. The readers' habits stay behind: the app where they open your work, the Notes feed that introduced them, the recommendations network that grew you, the subscription flow they've been trained to complete inside Substack's checkout. The export moves the address book. It doesn't move the attention. My friend could leave tomorrow with eight hundred email addresses and discover, as writers who have done it told her, that the audience's relationship was partly with the room, not with her. Which is the oldest fact in publishing, and the platforms are the new rooms.

The strongest case for Substack

The counterargument deserves its full weight, so here it is. For a writer with no audience, no technical skill, and no appetite for operations, ten percent of revenue buys a distribution system that individuals cannot build and would pay more than ten percent of their lives trying to. Email deliverability at scale is a genuine engineering problem. Payments, refunds, taxes, spam filters, an app, a podcast feed, a video pipeline, legal support through Substack Defender: assembled alone, this stack costs real money and real months. The writers who left for platforms charging two percent mostly kept their old readers and found new ones, and the writers who stay on Substack keep converting strangers. The deal is honest at the price quoted. I've also reviewed enough of this market to say that the fee is not the trap. The trap, if there is one, is the percentage of your attention the network asks for once it has you.

My own incentives aren't clean either. I thought about starting a newsletter when I quit consulting, ran the crossover arithmetic, and chose the multi-author site that pays this essay instead, so read this as a man reviewing a rival's kitchen.

The share worth watching

It depends, and the dependency is where a writer's growth comes from. If your new subscribers arrive because people go looking for you, you're using Substack, and the fee is maintenance. If the majority of them arrive through the app's recommendations, the app is using you, and the fee is rent. Watch the same number the company reports: the share of your growth that comes from its network. While that share climbs, the ten percent is cheap. The day it starts to fall, the conversation to have is not with Substack. It's with the arithmetic.