Somewhere tonight a person with half your knowledge is hitting publish on a paid post, and strangers are paying to read it. They were no more qualified than you. They simply put a price on the thing you keep giving away in group chats and long replies at work.
Meanwhile your own words earn your employer money. You write the reports, the explainers, the emails everyone forwards, and your salary stays where it was. Every review season someone else decides what your thinking is worth. AI can now produce a passable generic summary of anything in seconds, so bland writing loses value every month while a voice readers trust gains it.
Readers only pay for a handful of newsletters each. The writers who claim a subject early become the default name for it, and the ones who start later have to pry those readers loose from someone else's habit.
Here is the honest maths. Picture a reader reaching the end of your Tuesday email and paying $5 a month, the platform minimum. Substack takes 10%, and Stripe takes 2.9% plus $0.30 on the card and 0.7% for billing, so $0.98 goes and $4.02 stays, an effective rate of 19.6%. Plan your life around that figure. Starting costs you nothing, because Substack only charges once a reader pays. A publication that works sits somewhere in the $500 to $10,000 a month range, usually after one to three months of steady publishing.
Tonight, open the folder where your unfinished drafts already sit, pick out the ones a reader would have paid to read, and write down what they have in common.
That ease is also why this side hustle is so badly understood. Setting up takes you an afternoon, and that tells you nothing about how hard the business is. Whether your paid newsletter works comes down to a very specific piece of arithmetic: the fee stack, the rate at which paying readers leave you, and the number of hours you can keep giving it, week after week, indefinitely. Most people who start one never run that arithmetic, and most of them stop within a year. I would like you to be one of the ones who runs it first.
So this page runs it with you. Every fee, threshold and rule below comes from a source you can click at the bottom, and wherever a number is my own estimate rather than a published figure, I have labelled it as such.
Substack gives you a publishing tool, a mailing list, a website on a subdomain, an app presence, and a payments integration through Stripe. Publishing is free however big your list gets. Fees only begin when you switch on paid subscriptions.
Two details matter far more than they first appear to.
The minimum prices are fixed by the platform: $5 per month and $50 per year, per Substack's pledge documentation.
At $5 a month, each paying reader is a modest number with a real name attached. Roughly a hundred of them takes you toward $500 before fees and tax. That could be your car payment, covered by people who chose to read what you write each week, with every renewal still yours to earn. Price above the minimum if your subject can carry it. Who are the first ten people you can name who would pay $5 to hear from you every week?
Three separate fees come out of every payment you receive, and each behaves differently.
Here is what that does to real prices, on a US Stripe account.
The gap between 13.9% and 19.6% comes almost entirely from that flat $0.30, which is charged once per transaction whatever the size. A monthly plan gets charged twelve times a year. An annual plan gets charged once. Compare the same gross revenue: $10 a month nets you $100.08 over a year, while a single $120 annual charge nets you $103.38. The annual plan hands you $3.30 more per subscriber per year for the same money from the reader, purely because you paid the flat fee once instead of twelve times.
That is the first structural lesson here, and you will rarely hear it repeated. Annual billing matters less for cash flow than people think. Its real gift is that you pay a fixed fee once a year instead of twelve times. Are you offering your readers an annual plan, and making it the easy choice?
Substack's fee page quotes US Stripe rates. Your actual Stripe fees depend on the country your Stripe account is registered in, and they can be very different. Stripe's UK pricing page lists 1.5% + 20p for standard UK cards, 2.5% + 20p for EEA cards, and 3.15% + 20p for international cards, plus 2% for currency conversion.
If you live in the UK and most of your readers are American, you are paying a noticeably worse rate than you would with a British audience, and nothing in the Substack interface will tell you.
Substack offers localised pricing in CAD, GBP, AUD, EUR, USD, BRL, CHF, DKK, MXN, NOK, NZD, SEK and PLN. Its documentation says plainly that with localised pricing turned on "you may incur a slight extra fee on your Stripe transactions as Stripe charges 1% to 2% for currency conversion and settlement", and that Substack rounds converted prices upward to make up for it. It also notes that converted prices are refreshed only "roughly every three months based on exchange rate fluctuations". Between refreshes, your real price in each currency drifts with the exchange rate.
Substack's refund documentation contains a warning worth memorising: "Publishers should issue all refunds through the Substack Subscribers dashboard. Refunds issued through the Stripe dashboard will include Substack's 10% fee."
The Publisher Agreement backs this up: "all Revenue Share payments are non-refundable." Refund a $100 annual subscriber through the wrong dashboard and you hand $100 back to the reader while Substack keeps its $10. You end up $10 down on a sale that made you nothing. Write "refund from Substack, never Stripe" on a sticky note and put it on your monitor.
Look at that $4.02 again and multiply it by the readers already on your free list. A few hundred of them paying turns into a car payment covered, or the gym membership you cancelled to save money, or a weekend away you book without moving savings around. Small monthly payments arrive whether you had a good week or a bad one.
So you know what a $5 subscription is worth to you once the fees come out. When that money actually reaches your bank is a separate question.
Money moves from your reader to Stripe to your bank. Substack states that payments "usually arrive in your bank account within 48 hours of each transaction", though Stripe sets the actual schedule and new accounts usually start on a longer initial hold. There is no Substack-side minimum payout threshold, no invoicing, and no waiting for a monthly cycle.
Nothing is held back for tax. Every payout reaches you gross. The whole tax bill sits with you and arrives later, and that is the single most common nasty surprise in year two. If your first good year brought in a few thousand, would you have the tax portion sitting untouched when the bill came?
You can see your earnings in three places: the Stripe dashboard, the Substack earnings page at your publication's stats URL, and the Substack apps. Substack notes that the Stripe view "may include activity from sources other than Substack if you use your Stripe account for multiple projects", which is why both Substack and Stripe recommend you keep a dedicated Stripe account for the publication.
A subscription creates many small charges from the same card, and Stripe's documentation flags a specific consequence: "when an account owner disputes multiple payments within a recurring subscription plan, their bank creates a single dispute for the total amount against one of the charges." So a reader who forgot they subscribed and disputes twelve months of $10 charges can hit you with one $120 dispute, plus the dispute fee, against a single transaction.
Card networks typically allow 120 days to dispute. You get roughly 7 to 21 days to respond, the issuer takes 60 to 75 days to decide, and the whole cycle runs two to three months. All that time, the funds are held.
The defence is unglamorous and it works: a statement descriptor your readers recognise, renewal reminder emails, and a way to cancel that does not require emailing you.
That is one subscriber, priced properly. Whether your total ever grows depends on how many of those readers are still paying you a year from now.
Everything above is arithmetic you do once. Churn is arithmetic you do forever, and it is what really decides whether your paid newsletter works.
Think of your paid newsletter as a leaky bucket. Every month some of your paying readers leave, a few on purpose and some because a card expired. Your income is the number of people still subscribed this month. Everyone who ever subscribed and then left counts for nothing today.
Recurly's published churn benchmarks, updated with July 2026 data, put the Digital Media and Entertainment category at 4.14% total churn, split 2.55% voluntary and 1.59% involuntary, against an all-industry average of 3.60%. Recurly's page does not say clearly whether these are monthly or annual rates; the industry convention for consumer subscriptions is monthly, and an annual figure of 4.14% would be implausible for this category, so treat them as monthly.
Take 4.14% a month and compound it. After twelve months, roughly 60% of a cohort is still with you. You lose about 40% of your paying readers every year and have to replace them just to stand still.
At 500 paid subscribers, that means finding about 21 new paying readers every month, forever, before you grow by one. At 1,000, it is 41 a month. The bigger you get, the faster the treadmill runs. Could you find 21 new paying readers every single month, including the months you are tired or travelling?
Voluntary and involuntary churn are different problems
Roughly 38% of that churn in the Recurly data is involuntary: expired cards, insufficient funds, issuer declines. Nobody chose to leave you. The payment simply failed. This is the cheapest churn to fix, and solo publishers ignore it completely, because it happens silently inside Stripe.
Voluntary churn is about your content and the expectations you set. Substack's Retention dashboard shows you paid cohort analysis at 30 days, 6 months and 12 months, plus a paid churn chart and an unsubscribe reasons list for paid publications. That unsubscribe reason data is the most useful thing on the platform, and almost nobody reads it. Make a habit of reading yours once a month.
The annual renewal cliff
Substack's own documentation warns of "a sharper drop at 1 year as yearly subscriptions hit their renewal date". Annual plans hide churn for twelve months and then show it to you all at once. Your publication can look healthy in month nine and lose a third of its revenue in month thirteen. If you sell annual plans, your first honest read on retention comes roughly fourteen months after launch.
The dormancy rule is a publishing obligation
Substack's refund policy states it will issue a refund if "a publication is dormant", defined as inactive for at least six months for yearly subscribers or at least one month from the date of payment for monthly subscribers. Inactivity means no posts published.
Read that as a publishing schedule you have signed up to. Miss a month with monthly subscribers and Substack will honour refund requests over your head. You get no holiday, no burnout window, and no quiet quarter. This is the clause that ends most paid newsletters, and none of the launch guides mention it. Think about your hardest month last year: would you still have published?
Substack also honours refund requests submitted within 7 days of payment unless a subscriber perk has been claimed or you have marked the subscription non-refundable.
Free-to-paid conversion: what is actually known
Keeping paying readers assumes you found some in the first place. This is the step where one of your free readers decides to reach for a card.
You will see "5 to 10% of free subscribers convert to paid" quoted everywhere, including in the previous version of this page. Substack does not publish a platform-wide free-to-paid conversion rate. I could not find that figure in any Substack help article, in the Publisher Agreement, on Substack's growth features page, or in any first-party post. It gets repeated because it gets repeated, and I would rather you planned on something firmer.
Here is what can actually be sourced:
- Substack's own growth features page states that "The Substack Network drives 25% of paid subscriptions across the platform" and that readers on Substack are "3x more likely to pay for a subscription to your publication".
- The Reuters Institute's Digital News Report 2026 finds that across its basket of 20 tracked countries, "the percentage of people paying for news stayed roughly the same at 17% (compared to 18% in 2025)", with "10-20% appearing to represent a ceiling in most markets". In the UK the figure is around 10%.
- That report also notes that "in markets like Canada (29%) and the United States (21%) above-average percentages of people pay individuals", which is exactly the behaviour your paid newsletter depends on.
So across developed markets, roughly one reader in six pays for written journalism at all, and paying an individual writer like you, rather than an institution, is a minority habit inside that minority.
My estimate, clearly labelled as an estimate: for a general-interest newsletter with an engaged free list, sustained conversion in the 1% to 3% range is a realistic number to plan with. Rates above 5% do happen, but they cluster in publications where the content has a direct commercial payoff for the reader, which is a different business from writing. Plan on 2%, be delighted by 4%, and treat any launch-week spike as a one-off drawdown of goodwill from people who already like you rather than a rate you can repeat. At 2%, how big would your free list need to be before the paid side covered even one bill you care about?
Discovery and network effects, and where they stop
The real advantage Substack gives you over a self-hosted newsletter is distribution. Recommendations, the app, Notes and the leaderboards move readers between publications, and some of them will drift your way.
Substack's own numbers, from its February 2024 post on the recommendations network: "Recommendations, along with the Substack app, powers a network that drives 50% of all new subscriptions and 25% of new paid subscriptions on Substack." The current growth features page states the paid figure as 25% and adds that "more than 25% of subscriptions originate from our apps".
Notice the gap. The network brings in half of all new subscriptions but only a quarter of new paid ones. Readers who arrive through the network convert at roughly half the rate of readers who came to you directly. That makes sense when you think about it: ticking a recommendation checkbox at the end of someone else's signup flow costs nothing and says very little.
Substack's own figures have also shifted. An earlier Substack post put the network at "more than 40% of all subscriptions across the platform, and 12% of paid subscriptions". Two Substack posts, two different paid figures. I have not found a methodology note that reconciles them, so treat 12% to 25% as the plausible band for network-driven paid subscriptions, and plan from the lower end of it.
Leaderboards reward what you already have
Substack runs two leaderboard types per category. Top Bestsellers ranks by Annual Recurring Revenue and updates daily. Rising ranks by recent paid subscription growth and updates every few hours. Country leaderboards exist for the UK, France, Italy, the Netherlands and Spain, with your placement set by your Stripe account country.
Bestseller favours whoever is already winning: it ranks by revenue, so it can only ever show publications that already have revenue. Rising is the only realistic way in for a new publication like yours, and it rewards a burst of paid signups over steady quality. Your category choice is a request that Substack may override; it "reserves the right to adjust inappropriate assignments".
What the network cannot do
Recommendations are reciprocal, which turns them into a barter market where list size sets the exchange rate. A publication with 100,000 subscribers has no reason to recommend yours at 400. Substack's own case study on its growth page quotes Lenny Rachitsky saying "78% of new subscribers are now coming from other Substack newsletters recommending my newsletter. And 11% of paid." That is a writer with a 625,000-plus list describing the returns available to a writer with a 625,000-plus list.
The network moves readers who are already on Substack. It cannot create demand for your subject. If nobody wants to pay for weekly commentary on your topic, a leaderboard will not change that. Be honest with yourself: has anyone ever asked you to explain your subject, without you offering first?
Readers who pay for weekly commentary on your subject are recognition you can count. When a former colleague forwards your post to their team calling it the clearest explanation they have read, or a cousin asks how you got strangers to pay, you have proof your thinking is worth money. Write the next issue around a question someone on your list actually asked.
The recommendation network rewards writers who are already there when a reader goes looking. Every month a new publication on your subject starts collecting the recommendations, the Notes followers and the leaderboard spots that could have pointed at you. Readers settle into their few paid subscriptions and stop shopping. Each month you hold off, more of them have already picked another writer.
Substack Boost: growth that comes out of your margin
Boost is Substack's automated discounting system. Its documentation describes automatic discounts, free trials and offers "at the right moment to maximize your revenue". The upsell email carries "a default 20% discount for one year", and the multiple-subscription discount offers "up to 20% off".
The giveaway is in the payouts help article, which lists Boost first among reasons "your payouts are lower than you expected", and explains how to switch it off. That wording suggests Boost is on for you unless you turn it off.
Run the numbers on a nominal $10 per month subscription sold at a 20% Boost discount:
- Your reader pays $8.00
- Substack 10%: $0.80
- Stripe card: $0.232 + $0.30 = $0.532
- Stripe Billing 0.7%: $0.056
- You receive $6.61
Against a $10 headline price, you keep 66%. Whether that is a good trade depends entirely on whether that discounted reader would have paid you full price anyway. Substack's system optimises for conversion, and your margin is a separate goal it does not chase for you. Have you checked whether Boost is switched on in your settings right now?
It also muddles your reporting. Substack has a whole help article explaining why its Gross Annualized Revenue differs from Stripe's MRR: "Stripe considers the recurring value of those subscriptions to be their eventual full price rather than their current 33% off price. If the majority of your subscriptions came with discounts, the gap between what Stripe reports and what we report may be particularly large."
Two dashboards, two numbers, and neither of them is your bank balance.
Migration and list portability: read this before you launch
Substack's marketing leans on portability, and the Publisher Agreement supports it: "First and foremost, you own what you create", and your readers are described as "a list of subscribers that you control (your 'List')". The disputes article adds that "creators own their payment relationships, in addition to owning your intellectual property and mailing list. This gives you control and freedom to build your business on your own terms and leave Substack at any time if you choose."
That holds for your list. It holds a lot less for your revenue.
What you can take with you:
- A CSV of subscriber email addresses with subscription dates and plan types. Substack confirms "It's not currently possible to export a file with the full names of subscribers."
- Your posts.
- Your Stripe account, which is yours.
What you cannot take:
- The active paid subscriptions themselves, in any practical sense. Turning off payments means using the Danger Zone to "Disconnect Stripe", which Substack states will "Refund and notify all your paid subscribers by email" and "Cancel all paid subscriptions to your newsletter". Substack adds that "the balance you'll need in your Stripe account to successfully refund paying subscribers for subscription time they paid for but won't receive will be automatically calculated for you."
Read that slowly. Leaving means refunding every paying subscriber for unused time, out of your own money, and asking every one of them to type their card details in again on a new platform. Substack's own import guide describes the same friction from the other side: imported paid subscribers must be given a trial, and "after the trial ends, they'll be prompted to enter their payment information".
Ask anyone in the subscription business what happens when you ask a whole list to re-enter a card. You do not get everyone back. My estimate, and it is an estimate because no platform publishes this: expect to lose a substantial minority of your paying readers in a platform move, plausibly a quarter to a half depending on how much they care.
The two decisions that preserve your optionality
Use a custom domain from day one. It costs you $50 once. If you publish on a Substack subdomain for three years, every link, every search result and every citation points at an address you cannot take with you. Substack also states it may reassign an inactive subdomain or handle.
Keep a current export. Download your subscriber CSV regularly and keep it somewhere outside Substack. The Terms of Use reserve Substack's right "to terminate (or suspend access to) your use of Substack, or your account, for any reason at our discretion". If your account vanished tomorrow, when was your last copy of your list?
When the 10% stops being worth it
Substack's fee grows with you; the alternatives charge a flat price. Ghost's pricing page shows Ghost(Pro) tiers starting at $18 per month billed yearly and states that it takes no percentage of subscription revenue, with only payment processor fees applying. Beehiiv's pricing page shows a free Launch tier up to 2,500 subscribers and a Scale tier at $43 per month, and states it takes "0% of your paid subscription revenue".
Working out the crossover is simple: divide the flat annual cost by 0.10. Ghost's $29 per month tier costs $348 a year, which equals Substack's 10% at $3,480 of gross subscription revenue. Beehiiv's $43 per month tier costs $516 a year, matching it at $5,160. Above those points, on platform cost alone, Substack is the more expensive choice for you.
Two caveats, both important. Flat-fee tiers step up as your list grows, so the crossover moves with you. And the 10% is buying you the network, which the flat-fee platforms cannot sell you. The honest way to see it: Substack's fee is a distribution charge, and the question is whether the network is still bringing you meaningful growth at your size. If the Network tab in your stats says it is not, you are paying 10% for hosting.
Tax on subscription revenue
None of this is tax advice. Think of it as a map of the questions you will have to answer, with the current figures and where they came from.
Income tax and social contributions
United Kingdom. You must register for Self Assessment as a sole trader if you earn more than £1,000 in a tax year. For the 2026 to 2027 tax year gov.uk gives a Personal Allowance of £12,570, basic rate 20% from £12,571 to £50,270, higher rate 40% from £50,271 to £125,140, and additional rate 45% above that. For self-employed National Insurance, gov.uk gives a Class 2 rate of £3.65 a week for 2026 to 2027 with a Small Profits Threshold of £7,105, and Class 4 at 6% on profits from £12,570 to £50,270 and 2% above £50,270. Your subscription income is trading income, and the platform and processing fees are deductible business expenses.
United States. Self-employment tax is 15.3%, made up of 12.4% for Social Security and 2.9% for Medicare, and you must file if your net earnings from self-employment were $400 or more. The Social Security portion applies up to the annual taxable maximum, which the Social Security Administration set at $184,500 for 2026. Note that the IRS self-employment tax page still displayed the 2024 figure of $168,600 when I checked it, so go by the SSA announcement.
Form 1099-K. The IRS states that third-party settlement organisations must report "when the total amount of payments you receive for goods or services through the platform exceeds $20,000 in more than 200 transactions". The threshold has moved repeatedly in recent years, so check it each January instead of relying on last year's number. Either way, the IRS is explicit: "No matter the amount of reported payments, if you receive payments for selling goods or services, you must report all income on your tax return." So declare your income whether or not a form ever arrives.
VAT and GST on digital services: the part nobody mentions
This is where paid newsletters catch people out, and it follows directly from you being the merchant of record.
For business-to-consumer supplies of digital services, the place of supply is where your reader lives, rather than where you are. HMRC defines digital services as "e-services which are automatically delivered over the internet, or an electronic network, where there's minimal or no human intervention". An automated paid email subscription sits squarely inside that definition.
If you are UK-established and sell to EU consumers, HMRC's guidance gives you two options and no third: register for the non-Union scheme in one EU member state, or register for VAT in each member state where you supply consumers. There is no de minimis for a non-EU supplier. Your obligation begins at the first euro.
If you are EU-established, the European Commission's One Stop Shop site states: "Below this EUR 10 000 threshold, supplies of TBE (telecommunications, broadcasting and electronic) services and distance sales of goods within the EU may remain subject to VAT in the Member State where the taxable person is established." Below €10,000 of combined cross-border B2C sales you can charge your home-country VAT. Above it, you charge the customer's country's rate and file through the Union scheme OSS, quarterly.
Domestic UK VAT is separate again. gov.uk gives the registration threshold as £90,000 of taxable turnover on a rolling twelve-month basis, with a forward-looking test if you expect to exceed it in the next 30 days. Most solo newsletters will never get near this. It does not remove the EU obligation, which has no threshold for UK suppliers. Do you know roughly how many of your readers are in the EU today? Your Stripe dashboard can tell you in a minute.
The zero-rate question that is worth real money
Since 1 May 2020 the UK has zero-rated electronic publications, including "books, booklets, brochures, pamphlets, leaflets, newspapers, journals and periodicals - including magazines". A paid written newsletter published on a regular schedule looks a great deal like an electronic periodical.
The exclusion is the part to watch: "If more than half of an e-publication is devoted to advertising, audio or video content, its supply will be standard rated." If your Substack is mainly a podcast, or mainly video, or heavy with sponsors, it falls outside the zero rate. If it is mainly prose, it plausibly falls inside.
There is real tension here between two descriptions of the same product: an electronically supplied service with minimal human intervention, and a zero-rated electronic periodical. Which one applies decides whether you owe 20% on UK sales, and the answer depends on the specific mix of your publication. If you are getting close to a VAT registration threshold anywhere, this is worth an hour of a qualified adviser's time, and possibly a non-statutory clearance from HMRC. Please do not guess it from a blog post, including this one.
Practical mechanics
Substack integrates with Stripe Tax. Its documentation says the default tax category is "General - Electronically Supplied Services", that Stripe Tax "monitors your subscriptions so you'll know where and when you need to register", and that Substack "will only collect sales tax in locations where you've added a tax registration". To turn the integration on, you ask Substack support to enable it for your publication. Stripe Tax costs 0.5% per transaction in registered jurisdictions.
The trap: your listed price includes the tax until you register. If you cross a threshold and only then start collecting, every earlier sale had the tax buried inside a price you have already spent. That bill comes out of your margin, after the fact.
United States sales tax. Stripe's documentation notes that nexus can be economic as well as physical, that thresholds differ by state, and that "Stripe only monitors if you've reached an economic nexus tax threshold for sales outside of the region your business is based in". Because you are the merchant of record on Substack, and you are selling directly rather than through a marketplace, do not assume marketplace facilitator laws shift the collection duty away from you. Check it for your states.
Australia and Canada both run registration regimes for non-resident suppliers of digital services, each with their own turnover thresholds. I was not able to retrieve the current thresholds directly from the ATO or CRA while writing this, so I am not printing a number. Both authorities are linked below; confirm the current figure with them.
Consumer subscription law is tightening
United Kingdom. The Digital Markets, Competition and Consumers Act 2024 creates a new subscription contracts regime. The government's consultation response states: "We will legislate when parliamentary time allows and we anticipate that the regime will commence in spring 2027." It has already slipped twice, from spring 2026 to autumn 2026 to spring 2027. When it lands, you as the trader must give clear pre-contract information, send reminder notices before trials and long contracts auto-renew, allow online exit if signup was online, and honour two separate 14-day cooling-off windows with proportionate refunds.
United States. The FTC's amended Negative Option Rule, widely known as the click-to-cancel rule, was vacated in full by the Eighth Circuit on 8 July 2025 in Custom Communications, Inc. v. FTC, on procedural grounds, days before its compliance deadline. The FTC has since restarted the rulemaking. The Restore Online Shoppers' Confidence Act still applies, and state automatic renewal laws still apply. The specific federal rule is gone, and your legal exposure remains.
For you as a solo publisher, this mostly means: make cancelling easy, send renewal reminders, and never build your income on people forgetting they pay you. Would you be comfortable if every one of your readers knew exactly when they would next be charged? If yes, you are already on the right side of all of this.
Why most paid newsletters do not clear minimum wage
Now for the part the launch guides skip. I know it is uncomfortable, and I would still rather you saw it now than in year two.
The UK National Living Wage is £12.71 an hour for workers aged 21 and over from 1 April 2026. The US federal minimum wage is $7.25 an hour, unchanged since 24 July 2009, though many states set considerably higher rates.
A serious paid newsletter will eat far more than your weekends. Research, writing, editing, subscriber support, refund handling, admin and promotion for a weekly publication of any depth runs to 12 to 20 hours a week for most people. Take 15 hours a week as a mid-point. That is 780 hours of your year.
The UK case. 780 hours at £12.71 is £9,914 of gross pay-equivalent, before income tax and National Insurance. A £8 per month subscription paid by UK card nets you about £6.82, or £81.84 a year. To reach £9,914 you need about 121 paying subscribers, held continuously, with nothing left over for equipment, software, accountancy or research costs.
The US case. 780 hours at $7.25 is $5,655. A $10 per month subscription nets you $8.34, or $100.08 a year, so you need about 57 paying subscribers to clear the federal floor. At the $5 platform minimum, netting $48.24 a year, you need about 118. Against a $17 state minimum wage the same 780 hours costs you $13,260, which needs about 133 subscribers at $10 a month.
Those subscriber counts sound within reach. In practice they are much harder, and here is why.
First, churn. Holding 121 subscribers continuously at roughly 4% monthly churn means winning about five new paying readers every single month, indefinitely. So the real target is 121 signups plus about 60 replacements every year after that.
Second, the funnel above it. At an estimated 2% conversion, 121 paying subscribers implies a free list of around 6,000 engaged readers. Building an engaged list of 6,000 from a standing start, on a subject people will pay for, typically takes years. You are not paid for that build time.
Third, discounting. If Boost is on and a chunk of your readers are at 20% off, every number above gets worse by roughly a fifth.
Fourth, the hours can only go up. Publishing weekly through a month of illness or holiday is what Substack's dormancy refund rule asks of you for monthly subscribers. Treat those hours as the floor you must always clear.
So here is the question worth sitting with: if this newsletter earned you minimum wage for the next two years, would you still want to write it?
Substack's own about page states the platform has "5 million paid subscriptions and counting". Various third-party statistics sites quote much higher current figures. I could not verify any number above 5 million from a Substack source, so 5 million is what I am giving you.
Substack does not publish a median, a distribution, or the number of publications sharing those subscriptions. Any "average Substack earns X" figure is therefore built by dividing one published number by an unpublished one, and it would be meaningless anyway, because a few big winners drag the whole distribution far to the right.
Press Gazette's ranking, published in early January 2025, identified at least 52 newsletters earning $500,000 or more a year, collectively at least $40.2 million. Their method is worth understanding because it shows how little you can actually see: they infer revenue from Substack's public subscriber bands, where "hundreds" means more than 100 and "thousands" means more than 1,000, multiplied by the annual price. They state their figures "do not take into account the 10% taken by Substack as payment for the service" and "do not take into account subscription discounts, which are widely used and will cut into the actual revenue earned".
So the only public earnings data on Substack is (a) a lower bound, (b) before fees and discounts, and (c) limited by design to publications already big enough to appear in a band. It tells you a great deal about the top and literally nothing about the middle, which is where you would actually be.
The Reuters Institute notes individual creators with very large followings, including Heather Cox Richardson with "more than 2.5 million Substack subscribers". That is a total subscriber count; it tells you neither how many pay nor what she earns. It is also roughly the audience of a national newspaper, held by one person, so please do not measure yourself against it.
The honest summary: a small number of paid newsletters are excellent businesses, a modest number are decent second incomes, and the large majority never cover the hours put into them. Nobody publishes the shape of that distribution, and the platform has little reason to.
The arithmetic that has to be true before you start
You have now seen the fees, the churn and the ceiling. Here they turn into a sum you can run on paper before you give a single evening to writing.
Start from the income you want and work backwards.
- Decide the annual income you need this to produce. Be specific.
- Divide by the net-of-fees annual value of one subscriber at your intended price. Use the worked examples above; the headline price will flatter you.
- That gives you your required standing subscriber count.
- Multiply by about 0.4 to get the number of replacement subscribers you must find every year just to stay level.
- Divide the standing count by your conversion assumption, 2% unless you have your own data, to get the free list you need.
- Estimate honestly how long building that list will take you, and treat every one of those months as unpaid.
If that calculation produces a number you would turn down as a job offer, working harder on the newsletter will not save it. Change the price, change the subject, or add a second revenue line that does not depend on subscriber numbers.
Run the arithmetic and you can see what $10,000 a month, the ceiling here, would demand of you. Held for a year from subscriptions plus a second revenue line, a number like that lets you resign with months of living costs saved, and most newsletters never reach it. Set your price and your subject so the maths can work at your size. What number did your own sum just give you, and does it look like a job you would take?
Structural decisions that materially change the outcome
- Price above the minimum. The flat $0.30 per transaction makes the $5 monthly tier the worst-value plan on the platform, at 19.6% effective cost. If your writing justifies $8 or $10, charging $5 mostly hands the difference to Stripe, and your readers barely notice the saving.
- Push annual plans on purpose. They cut the flat-fee drag from twelve charges to one, and they delay churn. Just remember they also delay you finding out about churn.
- Own your domain from day one. $50 is the cheapest freedom you will ever buy.
- Fix involuntary churn first. It is roughly 38% of the total in the Recurly benchmark and asks nothing creative of you.
- Decide about Boost on purpose. It trades revenue per reader for more conversions, and Substack's payouts documentation names it first among reasons your money is lower than expected.
- Read your unsubscribe reasons. Substack collects them for paid publications and almost nobody looks.
- Keep a separate Stripe account. Both Substack and Stripe recommend it, and it makes your tax reporting manageable.
- Set aside tax from every payout. Nothing is withheld for you.
Delay here is paid for in words you give away. Another month of writing for free in comment sections and work emails, another month where your best ideas build somebody else's audience. You do not need a perfect launch. Pick a price above the minimum, write the welcome email tonight, and schedule the first post for this week. Readers cannot pay for a newsletter that does not exist yet.
Who should skip this
Be blunt with yourself about every one of these.
- You need income inside six months. The unpaid list-building phase is the business, and there is no version that skips it.
- You cannot publish on a fixed schedule indefinitely. Substack's dormancy refund rule turns a one-month gap into a refund event for monthly subscribers. If your life cannot make room for that, this is the wrong model for you, and a one-off product is a better fit.
- Your subject has no recurring need. People pay subscriptions for recurring value. A topic you could cover fully in ten excellent posts belongs in a book or a course.
- You are unwilling to be identifiable. Stripe requires verified identity details, and Substack warns that your support phone number, email and address may appear on receipts, invoices and card statements. A virtual mailbox helps. Full anonymity does not survive taking money.
- You will not do the admin. Refunds, chargebacks, VAT registration, list hygiene and Self Assessment are part of the job itself.
- You are building an asset to sell. You cannot move active paid subscriptions off Substack without refunding them and winning them back. The list has value; the recurring revenue is far less portable than it looks.
- Your paid content would count as regulated advice. If what sits behind your paywall amounts to regulated financial advice, investment recommendations, legal advice or medical advice where you live, charging for it does not exempt you from the relevant regulator. Check before you take a single payment.
- You only want to write. The writing is perhaps half the work. The rest is running a small subscription business with international tax exposure.
Where it does work
This model earns its keep under a narrow set of conditions, and they are worth naming because they run against most of the generic advice you will read.
- Your subject has a commercial payoff for the reader, so a $10 monthly charge sits against a benefit they can put a number on. That is why finance, professional and industry newsletters dominate the visible top of the leaderboards.
- You already have an audience somewhere else and are converting it rather than building from zero. That shrinks the unpaid phase from years to weeks.
- Your content is genuinely hard to get elsewhere, through your access, expertise or original work. Commentary on freely available news is the most crowded and least defensible category on the platform.
- Your newsletter is one revenue line among several, alongside consulting, sponsorship, speaking or a product, so subscription income never has to carry the whole load alone.
Count how many of those are true for you today. If two or more are, the arithmetic on this page can work for you. If none are, no amount of posting will fix it, and it is kinder to know that before you spend your evenings finding out.