Subscriptions are the most stable creator income there is. People pay monthly for access to a newsletter, a members’ feed, early releases, a community. The money then arrives whether or not this month’s post did well.
They are also the route where published expectations sit furthest from reality, by roughly a factor of ten in one direction and a factor of three in another. This page exists mainly to correct both, because getting them wrong costs people years.
The number nobody quotes#
beehiiv publishes an annual analysis of its own platform data, drawn from thousands of publications across 2021 to 2026. In the 2026 edition, the median conversion rate from free subscriber to paying subscriber is 0.62%.
Six paying subscribers per thousand.
The top quartile sits between 2% and 5%. By subject, sports newsletters average 1.93% and economics 1.28%, both well above the median. The top tenth of publications in each category range from 5.55% in technology to 30.80% in economics, with finance and investing at 20% and 18.69%. So the gap between typical and exceptional is enormous, in every single category.
Meanwhile the figure most people are working from is 5% to 10%, which Substack popularised years ago. That range describes its best performers. It was never a typical outcome, and planning with it will make your projections roughly ten times too optimistic.
The default price of a paid newsletter in 2026 is $10 a month.
beehiiv’s own platform data, which makes these the firmest figures on the site. Sources: beehiiv, The State of Paid Newsletters 2026, Press Gazette on newsletter pricing and churn, Humblytics benchmarks.
What that means for you#
At the median rate and the standard price, a thousand subscribers produces about $62 a month. To reach $2,000 a month on the same assumptions you need somewhere around 32,000 subscribers.
That is the answer to “how big does my list need to be to live off this”, and it is a great deal larger than the number in most people’s heads. Doing better is possible, and it comes from being in a high-converting subject, having an unusually engaged list, or charging considerably more than $10. It does not come from working harder at a 0.62% conversion rate.
The corollary is worth sitting with. With 2,000 subscribers a subscription produces beer money, while a digital product sold to the same list produces more, sooner. Subscriptions become the best route once the list is large. They are a poor first choice when it is not.
The other half of the problem: churn#
Everything above is about getting people to pay. Roughly half the difficulty is keeping them, and that half gets a fraction of the attention.
Monthly churn benchmarks for 2026, by subject:
| Subject | Monthly churn | Median subscriber lifetime |
|---|---|---|
| Food and drink | 5.06% | ~20 months |
| News | 5.47% | ~18 months |
| AI | 13.33% | ~8 months |
| Money and finance | 16.67% | ~6 months |
Paid communities generally run 5% to 10% monthly, with 4% to 5% considered excellent. Median lifetime value ranges from about $83 in community-shaped products to $230 in investing.
Compounding does the rest. At 5% monthly churn about 54% of a cohort is still there after a year. At 17%, about 11% is. Same product, same price, same effort, one business is building and the other is refilling a bucket.
Which produces an awkward conclusion about finance#
The standard advice is to write about money, because finance converts best. It does: top-decile finance publications convert at 20%, against 5.55% for technology.
It also churns worst. Money and finance shows the highest monthly churn of any vertical measured, at 16.67%, with a median subscriber lifetime of about six months.
Put those together and the picture changes. A finance newsletter converts a great many people who then leave within half a year, which means you are permanently reacquiring, and your growth has to outrun a leak that empties the tank annually. A food newsletter converts far fewer people who stay for twenty months. Depending on how hard acquisition is for you, the second can be the better business despite the worse headline number.
The general point is more useful than the specific one: conversion rate on its own tells you almost nothing. Conversion multiplied by subscriber lifetime is the number that decides whether this works, and nearly every published comparison quotes only the first.
Sources: Press Gazette on 2026 retention and churn, SubJolt churn benchmarks, Kourses on member retention. Checked 7 September 2026.
Annual plans are the single biggest lever you have#
Annual subscribers churn dramatically less than monthly ones, for the mundane reason that they only get the chance to leave once a year rather than twelve times.
The effect on the business is larger than it sounds. You collect twelve months of revenue up front, which solves the cash flow problem that makes creator income miserable. You remove eleven cancellation decisions. And you get a full year to demonstrate value rather than thirty days.
Practically: offer both, price the annual at ten months rather than twelve, and make it the default option on the page. If a meaningful share of your subscribers are not annual, that is the easiest improvement available to you and it requires writing nothing.
What goes behind the paywall#
The decision people agonise over, usually in the wrong terms. The question is not how much to give away. It is what someone is buying.
Three things reliably convert, and one reliably does not.
Utility. Something they use rather than read. The database, the tracker, the templates, the searchable archive. This converts best and churns least, because cancelling means losing a tool.
Completeness. The free version is useful and the paid version is the whole thing. The analysis, not the summary. The full data, not the chart. The long version.
Access. Getting to ask you things, or getting to the other subscribers. Expensive to sustain and it converts well, so price it accordingly and cap it before it eats your week.
Gratitude does not convert. “Support my work” is a donation model dressed as a subscription, and it produces a small number of generous people rather than a business. It also churns hard, because there is nothing to lose by cancelling.
On the split: the free tier is your entire acquisition channel. If nothing is public, nothing grows, and a paywall in front of everything is a slow way to stop. The publications that work tend to make the recurring, useful thing paid and the occasional, discoverable thing free.
Which platform, and why it matters more than the features#
The important difference is not the editor or the fee. It is whether the platform brings you subscribers.
Patreon has paid creators more than $3.5 billion in total, so the model plainly works. But around 94% of its subscribers arrive from a creator’s existing audience elsewhere. Patreon converts an audience you already have; it does not find you one.
Substack reports that over half of new subscribers come from inside its own network. Its recommendation system actively distributes, which no other platform here does, and it is strongest for writing, journalism and commentary.
So if you already have an audience elsewhere, platform choice is mostly about fees and features and you should not spend a week on it. If you do not, that difference is the only thing that matters, and it should decide where you start.
Sources: Substack’s own comparison, Everything-PR on Patreon.
Turning it on#
A launch that works is an explanation followed by a deadline, not an announcement.
The founding-member mechanic is standard because it does two useful things at once: it gets your first cohort in at a price that rewards early trust, and it gives you a genuine deadline rather than a manufactured one. Something close to this, sent to the free list:
From the 21st there’ll be a paid version of this newsletter.
What stays free: the Thursday piece, same as always. What’s new and paid: the full data behind each one, the searchable archive of everything since 2023, and the monthly breakdown I currently only do for clients.
It’s $12 a month, or $120 a year. For the first two weeks it’s $90 a year for as long as you stay subscribed, because you’re taking a bet on something that doesn’t exist yet.
If the free version is what you want, nothing changes and you don’t need to do anything.
Three things that message does. It says plainly what stays free, which prevents the resentment that kills open rates. It describes the paid tier as things rather than as support. And it reassures the majority who will not pay, who are still your acquisition channel and should not feel pushed out.
Then say it again a week later, and once more on the last day. Most of your list did not see the first email.
Keeping people, where the money is#
Given the churn numbers, retention deserves at least as much of your attention as acquisition, and it will not get it unless you make yourself look.
Watch cancellations from month one. Not as a vanity-adjacent metric but as the number that determines whether growth compounds. If churn is above 10% monthly you have a product problem, not a marketing one.
Ask people why they left. One line on the cancellation flow, or a short email a day later. Most will not answer. The ones who do will tell you the same two things over and over, and those two things are your list of what to fix.
The first month decides a lot. New subscribers who do not use the thing in the first few weeks cancel. Send them straight to the best of the archive rather than waiting for the next issue to arrive.
Publish on a rhythm you can sustain for years. The usual failure here is stopping, not converting badly. Subscribers notice a gap immediately and it is the cheapest possible reason to lose them. Monthly and reliable beats weekly and abandoned and you should choose the cadence assuming a bad month rather than a good one.
Price increases apply to new subscribers. Grandfathering existing ones costs you little and removes the single most common reason for a cancellation wave.
What changed in 2026#
Direct payment grew while platform advertising got harder, which shifted the balance towards subscriptions for anyone with an audience. YouTube doubled its Partner Programme thresholds and named three categories of content it will no longer monetise. Search referral traffic to publishers fell around 38%. Anything you own became worth more relative to anything you rent.
The platforms noticed and started competing for the same money. Meta launched paid subscriptions across Instagram, Facebook and WhatsApp in May 2026. YouTube took Premium Lite global with a 60% share of net revenue to creators, double the 30% on standard Premium.
None of that changes the conversion arithmetic or the churn arithmetic. What it means is that there are more places to run a membership, and the ones attached to a large platform bring their own discovery, which remains the scarce ingredient.
Who should not do this#
Anyone with a list in the low thousands, for the arithmetic reasons above. Build the list, or sell a product to it instead.
Anyone who cannot commit to a schedule for years rather than months. This route punishes inconsistency more than any other.
Anyone whose audience has no financial reason to pay. If the information does not save or make your readers money, or replace something they would otherwise buy, you will sit well below a 0.62% median rather than above it.
Anyone who needs income this year. Months to the first subscribers, years to a meaningful income, and unlike advertising it compounds only if you stay.
Where to start#
- Count your actual engaged audience. Multiply by 0.006, then by the price you have in mind. That is your realistic first-year monthly revenue. Decide whether that is enough before doing anything else.
- If the number is small, build a product instead and come back when the list is bigger. This is the right answer more often than not.
- If it is worth it, decide what paying gets you: utility, completeness or access. Write it down as a list of things, not as a description of support.
- Price at $10 to $12 monthly unless you have a reason to go higher. Offer annual at ten months and make it the default.
- Launch with a founding-member window and a real deadline. Say it three times over two weeks.
- Keep publishing publicly. The free work is the only reason anyone finds the paid work.
- Watch cancellations as closely as sign-ups from the very first month, and ask leavers why.
- Raise prices for new subscribers only.
The income target calculator runs the conversion arithmetic for a target figure. Owning your audience covers building the list, which is the constraint that actually binds.