Creator Agency Commission: How Agencies Actually Charge

Ask three agencies what they charge and you get three percentages within a few points of each other. Ask what the percentage applies to and at least one of them has to go and check the contract. The second question moves more money than the first, in one direction, every month, for as long as the deal lasts.

This page is about how a creator agency commission is actually built: the full stack of deductions between the fan’s card and the creator’s bank, and the handful of clauses that cost more than any percentage you could argue over. What follows describes how these contracts are commonly built. It is not legal advice on yours.

What does a creator agency commission actually apply to?

To whatever the contract names. Where it names nothing, the base is set in practice by whoever produces the statement. Three bases are in circulation, and they are not variations on a theme.

• Gross sales. The headline number on the platform dashboard, before the platform takes its own share.
• Platform net. What the platform actually transfers. This is net revenue, and it is the only base where every unit being divided has genuinely arrived.
• Net of agreed costs. Platform net minus named, pre-agreed expenses (paid acquisition, a shoot, refunds) before the split.

Alongside the base sits the scope: which sales the base is built from. Subscriptions, PPV, tips, custom orders, brand deals, and revenue from fans who were already subscribed at signature. Each of those needs a line. A single sentence about “all revenue” settles the argument only until the first month one of those sources gets big.

What does the full stack of deductions look like?

Seven lines, in order, each one taking from what the previous one left. Write them out once and the negotiation changes shape, because most disputes are about rank, not rate.

#	Line	Taken by	Applies to	Negotiable?
1	Price the fan paid	n/a	n/a	You set it
2	Platform cut	The platform	The fan’s payment	No: whatever your platform charges
3	Refunds and chargebacks	Fan, via platform	Sales already counted	No, but the timing is
4	Transfer and currency costs	Processor, bank	The amount moving	Partly, by payout frequency
5	Agency commission	The agency	Line 1 or line 2: this is the clause	Yes
6	Rebilled expenses	The agency	Named items only	Yes, and it needs a cap
7	Tax on what is left	The tax authority	The creator’s income	No

Two rows carry the whole document. Row 5 decides whether the commission is a share of money that exists or a share of a number on a screen. Row 6 is where a low rate goes to get its margin back. The full ordering, including outsourced chatting providers and chatters paid on sales, is laid out in revenue split.

How much does the base change what the creator keeps?

Enough to be worth more than the rate negotiation, and the size of the gap follows one rule.

Multiply your commission rate by the platform’s rate. That product is the share of gross that changes hands purely because of the wording of one clause. Nothing else about the deal changes. Nobody works differently. The same conversations produce the same sales, and a different slice of them lands on a different side.

That gap does not arrive once. It recurs every month the contract runs, so the honest way to look at it is annual: the monthly gap multiplied by twelve, then compared against whatever the agency offered as a concession on the rate.

Run it on your own figures rather than on anyone’s example. One closed month, seven lines:

Step	Where the number comes from	Your figure
A. Gross sales	Platform dashboard, sales for the month
B. Amount transferred	Platform payout statement, same period
C. Platform cut	A minus B
D. Commission on gross	A × your rate
E. Commission on net	B × your rate
F. Cost of the word “gross”	D minus E
G. Annual cost of one word	F × 12

If F is not zero and the contract does not say which of D or E is meant, you have found the next conversation to have.

What is the commission supposed to buy?

Everything required to produce the sales it is calculated on, and nothing that would have existed anyway. That single test resolves most rebilling arguments before they start.

Inside the commission, by default: the people writing the conversations, the shift coverage, the scheduling, quality control, reporting, and whatever software the agency runs. Those are the agency’s costs of doing business, and they are the substance of what a chatting team costs on the agency’s side of the line.

Outside it, and legitimately rebillable if agreed in advance with an amount attached: a shoot, travel, paid acquisition, a commissioned asset. The rule that holds up in practice: if the spend benefits the agency first, it sits inside the commission; if it benefits the creator first and would have happened without an agency, it can be rebilled.

A high rate covering real spend is cheaper than a low rate with an open rebilling line. That comparison is only possible when both are written down.

Which clauses cost more than the percentage itself?

Four, and none of them appear as a number. They are the ones to read first, because a rate is visible at signature and these are not.

1. Commission surviving termination. Bounded by a short duration, a rate that steps down monthly, and application only to fans acquired during the contract, or it is a tax on leaving.
2. Uncapped rebilled expenses. A cap, in currency, per month. Without it, row 6 of the stack is the real rate.
3. Exclusivity with no end date. Exclusivity is normal; open-ended exclusivity is a different product.
4. Unilateral rate changes. If the agency can revise the rate with notice and the creator cannot leave inside that notice, the rate is not a term of the contract.

A fifth, less obvious: no audit right. If the creator cannot see the underlying sales, the commission is whatever the summary says.

How does the commission structure change what gets sold?

It sets what the people writing the messages optimise for, which is the part no clause mentions. Pay structures reach the inbox faster than any training document.

Our data is blunt about the failure mode: a sale pitched before the sixth message converts roughly a third worse than one placed after about ten exchanges. Any arrangement that pays on activity (per pitch, per message, per shift target) manufactures exactly that behaviour, and the loss lands on both sides of the split at once. The same corpus shows the cheapest fixes are stylistic: ending a sales message with an ellipsis is the worst closer we have measured, and asking a closed question at the moment of closing costs several points of conversion. None of that is a commission problem until the commission pays for the wrong thing.

Structure the variable pay on tracked sales, hold the rate, and the incentive question stops. What that does to the whole model, against a retainer, is flat fee versus commission.

How do you check that the commission charged matches the contract?

Reconstruct one closed month from the platform’s own statements, not from the agency’s summary, and compare three figures: gross sales, amount transferred, and the base the commission was actually taken on.

• Do it source by source. A blended total hides the one source where almost everything dissolves.
• Compare like months. A gap that moves by several points with nothing on the payout statement to explain it is usually a deduction nobody wrote down.
• Watch the period, not just the amounts. A calendar month and a platform payout cycle are different windows, and a sale near the boundary lands in one or the other.

The point of the exercise is that both sides can run it and get the same number, which is the whole test of a commission clause worth signing. How that reporting habit fits the rest of the operation is covered in running a chatting agency.