Own video platform vs YouTube: what you actually keep

What YouTube, Patreon and your own site each keep

Three published sets of rates, side by side.

YouTube: three rates, not one

Creators quote the first rate and forget the other two. YouTube’s own announcement sets the split at 55% to the creator on long-form videos and 45% on YouTube Shorts. Subscription income moves differently: 30% of net Premium revenue and 60% of net Premium Lite revenue enters a pool shared across channels, and your slice of that pool is then split 55/45 by watch time and views. Fan funding pays best of all. Channel memberships, Super Chat and Super Thanks run at 70/30 in your favour.

Two conditions reduce the headline number. Videos using eligible commercial music share revenue with music partners, so a channel leaning on licensed tracks earns below 55%. And from February 2027, keeping Shorts revenue means holding 10 million qualified views across the trailing 90 days, checked monthly.

Patreon: one fee, plus the ones underneath it

Patreon’s published fee structure puts every Patreon page created after 4 August 2025 on a flat 10% platform fee. Underneath sit payment processing fees, which vary by currency: in sterling, 3.4% plus a fixed £0.35 per transaction, with a further 2.5% when a patron pays in a currency other than your payout currency.

Two Patreon deadlines get conflated and they aren’t the same thing. Legacy billing models end on 1 November 2026, when every creator moves to subscription billing. Legacy platform pricing is separate: older creators can keep their original rate, and the way to lose it is to unpublish and republish the page, which moves the account onto the standard 10%.

Your own checkout

One layer of fees: whatever your payment processing provider charges, which depends on the payment method, the card’s country of issue and your own market. Then a fixed monthly cost for the service itself, which changes the shape of the maths rather than the rate.

A £10 monthly subscription through all three routes, with processing at 2.9% plus £0.30 on the direct checkout:

Two things fall out of that table. Patreon on the web returns more than a YouTube channel membership at the same price, and a YouTube membership at 70/30 returns more than YouTube ads at 55/45. Creators who leave YouTube over a disappointing ad RPM sometimes haven’t checked that a better-paying option was sitting on the channel they already had.

The third column also carries a fixed monthly cost the other two don’t, which is why the rate on its own settles nothing.

The app-store layer nobody adds up

Every fee calculator you’ll find online stops at the table above. There’s a fourth layer, and on mobile it’s the largest one.

Sell access through in-app purchase on iOS and Apple takes 30% of the sale, dropping to 15% once a subscriber completes a year of continuous billing, or 15% from the start for members of the Small Business Program. That applies whether the app is yours or somebody else’s. Buy a Patreon membership inside the Patreon app and Apple’s share comes off first, with Patreon’s own fee applying on top, so two companies take a slice of the same payment before it reaches you. Users never see any of this. They see one price.

The rules now vary by market

Since May 2025, apps on the US App Store can show buttons and links sending users to an external purchase page, and Apple cannot charge a commission on what they buy there. In December 2025 the Ninth Circuit ruled that Apple should be allowed a reasonable commission covering its own costs, but nothing can be charged until a district court approves a rate. Apple filed its proposal in August 2026: 15% for standard apps, 5% for Small Business Program members. No rate has been approved at the time of writing.

App-store economics now vary significantly by market, so iOS revenue should be modelled separately for the US, the EU and other regions. The rules in each are moving at different speeds and in different directions, and a single global assumption will be wrong somewhere.

Why the commission rate is the least important number

This is the part that decides whether the project works, and most analysis skips it.

Two levers, the same audience

Take 50,000 monthly active viewers and a £10 membership.

At 1% conversion, that’s 500 members paying £10 on YouTube, of which you keep £7.00 each: £3,500 a month.

Move the same 500 members to your own checkout and you keep £9.41 each: £4,705 a month, before the platform’s fixed costs. A 34% improvement, and that is the entire benefit of the commission change.

Now hold the commission constant and move conversion from 1% to 3%. That’s 1,500 members at £9.41: £14,115 a month.

The rate gave you a third more. Conversion gave you three times more.

A lower commission on its own doesn’t build a case for anything. What builds it is audience size, conversion, pricing, retention and your ability to reach viewers directly. The rate is printed on a pricing page, which is why it gets compared; the other five are the actual business.

What conversion rates actually look like

Treat these as ranges to test, not facts to plan around. Substack’s published guidance to writers is 5% to 10% of free readers converting to paid, while independent collections of reported figures cluster nearer 3%, with most publications landing between 2% and 5%. For channel memberships the repeated band is 0.5% to 2% of active viewers, and that number comes from tooling companies rather than from a primary study. Niche matters: education and finance channels sit at the top of the band, entertainment and gaming at the bottom, where volume makes up the difference.

Nothing about moving to your own site raises conversion by itself. In the first weeks it usually drops, because you’ve added friction that YouTube didn’t have: leaving one page, creating an account, entering a payment method. What lifts it over the following year is what ownership makes possible.

Work out your own break-even

Six inputs. Twenty minutes in a spreadsheet.

The six inputs

1. Your real baseline. Open YouTube Studio and take your actual earnings across the past twelve months, split by source. Not an RPM benchmark from a blog post, your number. That’s what the new route has to beat.

2. Monthly active users, not subscribers. Your subscriber counter includes everyone who ever clicked it, including people who stopped watching in 2023. Conversion happens among users who watched you this month, and that’s a smaller and far more honest denominator.

3. Your conversion band. Pick a low case and a high case. If you’ve never sold anything directly to this audience, start at the pessimistic end.

4. Price and billing period. An annual plan changes two factors at once: one payment processing fee instead of twelve, and a renewal decision taken once a year rather than every month.

5. Your keep rate after all four layers. Web, iOS and Android separately, because they aren’t the same number, and iOS split by region. Weight them by where your audience actually watches.

6. The fixed monthly cost. This is your floor. Divide it by your margin per paying member and you have the number of members that gets you to zero. Everything past that is income.

If you’re weighing building the whole thing yourself against licensing software, input six is where the two paths diverge sharply. We’ve broken that comparison down in build it yourself or use a white-label platform and in what an OTT app actually costs to build.

The line item everyone forgets

There’s a seventh input, and leaving it out is the most common reason these projections don’t survive contact with the real world.

YouTube’s 45% buys distribution. Recommendations, search, the home feed, millions of users arriving without you paying for them. Your own site sells nothing to people who don’t already know you exist. Whatever you’d spend on ads, sponsorships or cross-promotion to replace that flow belongs in the model as a monthly line.

Which is why the version of this decision that works for most established creators isn’t a departure at all. Keep uploading to YouTube, treat it as the top of the funnel it already is, and move the paid relationship somewhere you control. YouTube stays the discovery engine. Your own service becomes the place where a viewer turns into a customer you can reach directly. In the £10 example above, that arrangement leaves roughly £9.41 per member before fixed costs, against £7.00 for the same member on YouTube, with the discovery still running for free.

What a service of your own has to cover

Moving the checkout means taking on jobs YouTube and Patreon were doing quietly. This is what to check before you commit.

Payment methods, plans and billing

Multiple payment methods, because a card-only checkout loses users in markets where cards aren’t the default. Recurring billing with retries and dunning, or you’ll lose members to expired cards rather than to churn. Annual and monthly plans side by side. Regional pricing if your audience is spread across the world. Payment security and PCI scope handled by the provider rather than by you, since that’s specialist knowledge you shouldn’t have to acquire to sell videos.

Better Media Suite runs six monetisation models against the same catalogue: subscriptions, rentals, purchases, advertising, free ad-supported channels and freemium. From what we see across the 30-plus tenants on the platform, operators who launch with one model and add a second within the first year tend to hold revenue steadier than those relying on a single membership tier. That’s an observation from working with those services rather than a measured result, and your own mix is worth testing. There’s a fuller breakdown in the monetisation models you can run side by side.

Apps on the television

A subscription that plays on the living-room TV tends to get used differently from one living in a browser tab, and habitual use is what keeps members paying.

The free route here closed some time ago. Roku retired Direct Publisher on 12 January 2024 and removed the channels built with it, so every path onto a television now runs through an SDK app or through a company that maintains those apps for you. Better Media Suite ships to more than ten connected TV platforms from one catalogue, which means uploading a video once rather than eleven times, and one set of features to keep consistent across every screen. Publishers making this same calculation are covered in publishers moving off YouTube.

The team behind the software

Streaming infrastructure fails in ways a website doesn’t: a certification rejection, a DRM licence error, a playback failure on one TV model. Ask who answers when that happens, how quickly, and whether support is a ticket queue or a named contact. Ask what happens to your subscriber data, billing records and account status if you leave. A provider that can’t answer the second question is selling you the same dependency you’re trying to escape.

When you should stay exactly where you are

Four situations where the arithmetic says don’t.

You can’t reach your audience off-platform. No email list, no Discord, no newsletter. Without a channel you control, you’re asking people to find a new website through a link in a video description, and conversion at that stage sits at the bottom of any reported band. Spend six months building the list, then run the model again.

Your content depends on algorithmic discovery. Broad entertainment, trend-driven formats, anything where most views come from users who’ve never heard of you. That audience arrives, it doesn’t follow. Your paying audience is the one that searches for you by name.

You have no members-only cadence. A subscription is a promise of exclusive content on a schedule. If everything you make goes public anyway, you’re asking people to pay for goodwill, which works for a while and then doesn’t.

You’re moving for safety rather than growth. Changing provider isn’t the same as owning your audience. Vimeo told sellers on Vimeo On Demand, which closes on 20 November 2026, that content and subscriber data wouldn’t be migrated and that any transition is a fresh start, then pointed them towards another product owned by the same company. If your reason for moving is that you don’t trust a vendor, confirm that the destination gives you bulk export of subscriber and billing records, or you’ll be having this conversation again in three years. Sometimes the better use of the year is the community instead, which we’ve written about in building a community around your platform.

Where that leaves you

The rate you pay is the easiest number to compare and the least important one in the outcome. What decides this is how many of your viewers will pay, at what price, for how long, and whether you can still send them a message next year without asking permission. Keeping YouTube as the discovery layer and moving the paid relationship somewhere you own is the arrangement that holds up, and it only starts paying once the audience and the offer are already there.

Better Media Suite gives you the checkout, the subscriber records, six monetisation models and apps across more than ten connected TV platforms, running from a single catalogue.

Want to see what your own numbers look like on a service you control? Request a demo and we’ll build the model with your actual audience data.