Fansite Revenue Breakdown: Where the Money Actually Goes

A fansite revenue breakdown is the difference between a business and a spreadsheet. Gross subscription revenue is the number people quote; what reaches the operator after processing, disputes, the creator split, and the costs that only appear at scale is the number that decides whether the model works. This walks the dollar down.

Start with the split, because it sets the ceiling

On the incumbent platform the headline is simple: OnlyFans takes 20% and the creator keeps 80%, on a business that Variety reported at scale. That single number is why operators look at building: 20% of a large number is a business, and the platform is not sharing it.

Running your own platform changes the question from “what cut do I pay” to “what cut can I keep after I pay for everything the platform used to do”. Those costs are real, and they are the rest of this article.

Payment processing costs more here than the rate card suggests

Adult content is a restricted category for mainstream processors, named in Stripe’s restricted businesses list, so an operator uses high-risk acquiring. High-risk pricing carries a higher percentage and a higher per-transaction fee than standard card processing, and it usually adds a rolling reserve: a percentage of your volume held for a period before it is released.

The reserve is not a cost, it is working capital. Operators model the rate and forget the reserve, then discover that a growing month ties up more cash than it releases. Model both.

Chargebacks are a cost and a threshold

Every dispute costs you the transaction, a fee, and a step towards a ratio you cannot cross. The card schemes run monitoring programmes with thresholds on dispute rate, and crossing them brings higher fees, mandatory remediation, and eventually the loss of processing. In this category the practical exposure is higher than most verticals because “I do not recognise this charge” is an easy dispute for a customer to file about a subscription they would rather not discuss.

That makes descriptor clarity, rebill notice, and easy self-serve cancellation into margin decisions, not UX niceties. Our chargeback management guide covers what the ratios have to stay under and which controls actually move them.

The creator split is a market rate, not a lever

An operator who tries to fund the business by taking a bigger cut than the market discovers the constraint quickly: creators move, and the ones who move first are the ones with an audience. In practice the split is set by what creators can get elsewhere, which means your margin has to come from operating costs, retention, and volume rather than from squeezing the split.

This is the single most common modelling error. A spreadsheet that gets to profitability by assuming a below-market split is not a plan, it is a prediction that your best creators will leave.

The costs that only appear at scale

Four line items that are invisible at ten creators and material at two hundred.

Media delivery. Video is the product, and bandwidth and storage scale with both catalogue size and viewership. Transcoding is a real compute cost.

Support. Fans contact you about billing far more than about content. Every thousand subscribers adds a predictable ticket load, and unanswered billing tickets convert into chargebacks, which is the expensive path.

Compliance. Age assurance is charged per check, so it scales with traffic rather than revenue, and the conversion cost of the check itself is real. Add identity verification per creator and the record-keeping obligations covered in the operator legal guide.

Payouts. Cross-border payouts carry per-payment costs and FX spread. At a small roster this is a rounding error; across hundreds of creators paid twice a month it is a line item.

Retention is the term that dominates everything else

Subscription businesses are decided by how long a subscriber stays, not how many arrive. Two platforms with identical acquisition can differ by an order of magnitude in value per subscriber if one holds them four months and the other holds them one.

The operator lever is not content volume, it is the reason to still be subscribed in month three. That usually means something delivered on a schedule, a back catalogue worth access to, or a community element that does not exist elsewhere. Volume alone produces churn, because a fan who can consume everything in a week has no reason for a second month.

Retention feeds back into the payment terms too. Acquirers price on risk, and a platform with long average tenure and low dispute rate negotiates better than one with churn and refunds. Retention shows up twice in the model: once as revenue, once as cost of processing.

Refunds and involuntary churn

Two leaks that sit outside most models.

Refunds are a policy decision with a cost attached. A generous policy lowers dispute rate, which protects processing, and raises refund cost, which is a line item you control. A stingy policy trades a controllable cost for an uncontrollable one, which is usually the wrong trade in this category.

Involuntary churn is the subscriber who wanted to stay and whose card failed. Expired cards, insufficient funds and issuer declines are a meaningful share of cancellations, and much of it is recoverable with retry logic and pre-dunning notice. An operator not running dunning is losing revenue from people who never chose to leave.

Both belong in the model from the start, because both scale with the subscriber base rather than with effort.

The numbers to instrument from day one

Four measurements make the model real rather than theoretical, and all four are much harder to reconstruct than to record.

Revenue per creator, by cohort month. Not total revenue, which is dominated by whoever joined most recently. Per creator by month tells you whether creators mature into earners or peak in week two.

Dispute rate, tracked weekly. Weekly rather than monthly because the schemes measure against thresholds and a bad month is easier to correct if you see it in week two. Track it by acquisition channel as well as in aggregate, since one bad channel usually accounts for most of it.

Involuntary versus voluntary churn. These have different fixes. Voluntary churn is a product and content problem; involuntary is a dunning and retry problem, and confusing the two sends you optimising the wrong thing.

Payout lag. Time between a creator earning and being paid. It is not a cost line, but it is the single strongest predictor of whether creators stay, which makes it an economic measurement even though it does not appear in the P and L.

Instrument these before launch. An operator who can answer all four after ninety days can make decisions; one who can only report gross revenue is guessing with more confidence.

What the model looks like from the operator’s side

Put together, the operator’s economics are: gross revenue, minus processing and reserve drag, minus disputes, minus the creator split, minus delivery, support, compliance and payout costs, minus the platform cost or the engineering cost of not having one.

The reason white-label exists is that most of those middle terms are fixed costs an operator would otherwise fund alone. A managed platform converts acquiring, compliance, delivery, and platform engineering into one predictable line, which is worth more than a lower headline percentage if the alternative is running them yourself. The comparison against a self-hosted script, where those costs come back to you individually, is set out in white-label versus building your own, and the pricing side is in fansite pricing strategy.

Model it with the reserve, the disputes, and the support load in from the start. A model that works only when nobody asks for a refund is not a model.

Wick runs payments, compliance, and delivery as one predictable cost under your brand, so the economics are legible before you launch. See Wick’s pricing