Your Agency Is Not Performing — What to Do in the Next Seven Days

Last updated August 21, 2026 · by the Perlage Studios team

This guide assumes you are past wondering. Something has been wrong for a while: flat months, thinner message income, reporting that arrives late or not at all. You have already read enough to suspect the agency. The diagnosis itself lives in the warning-signs guide; this is what comes after it. Seven days of work, most of it quiet, that ends with you staying, renegotiating, or leaving on evidence instead of a mood. Nothing here is legal advice, and almost none of it needs the agency’s permission to start.

What should I have in hand before I say a word to my agency?

Thirty days of your own numbers, pulled from your own dashboard before you say anything to anyone: net revenue by stream, subscriber counts, message income, posting frequency, and how long the agency takes to answer you — screenshotted, dates visible. Every item on that list comes from access you already have, so nothing requires the agency’s cooperation.

Days one and two are collection, not confrontation. The moment you raise performance, the story starts being managed, so gather the record while everything still looks the way it actually is. Pull the last 30 days plus the same window from the two months before: revenue split by subscriptions, messages, tips and posts, subscriber movement, what actually got posted and when. Screenshot it all with dates visible and keep the file somewhere the agency cannot reach.

Work from one consistent figure. OnlyFans takes 20% of each fan payment, so decide up front whether every number you write down is gross or net of that, and never mix the two in a comparison. Which split your agency’s cut is computed on is a separate question, and it comes up on day four.

If checking your own account through your own login feels like going behind their back, it is not. In Stage v. Unruly Agency LLC (California Labor Commissioner, Case No. TAC-52876), the record shows the creator kept her own OnlyFans access and used it to read the chats her account managers were sending and to track her revenue. Reading your own dashboard is the baseline, not an act of war.

Who is actually talking to my fans, and how do I check?

Assume it is not who you think. Reporting has documented chatters running three or four creators’ inboxes at once, and AI tools drafting replies while a human presses send. Read your own sent messages for an hour: tone breaks, recycled scripts, discounts you never approved. Whoever holds your inbox holds the biggest lever on your revenue.

Messaging is not a side channel. An agency founder quoted by Vice estimates creators earn 50 to 60 percent of their income through messaging, and the same reporting describes chatters assigned the inboxes of three or four creators at a time, often working from Pakistan, India or the Philippines. Fortune put the going wage at roughly $3 an hour, in shifts of up to 12 hours, six days a week. That is who typically holds the revenue engine of an account like yours.

The newer version is software. Fortune reported that ChatPersona signed over 110 agency clients within eight months and that FlirtFlow charges $1,000 to onboard a creator, with humans still pressing send, in Fortune’s telling, to get around OnlyFans’ prohibition on AI responses. Whether your agency runs anything like this is easier to see in your sent folder than to get answered in a meeting.

So day three is an hour of reading as a fan would. Open your sent messages across one full day, including the hours you were asleep. Look for tone that breaks mid-thread, the same script sent to different spenders, discounts or customs you never approved, prices drifting from your list. Then pick your five biggest spenders and read each thread top to bottom. In the Unruly determination, subscribers believed they were writing with the creator while account managers ran the conversation. Ask whether a fan reading your threads would say the voice is yours.

There is a commercial edge to this, beyond taste. Two subscribers have sued OnlyFans’ parent companies Fenix Internet and Fenix International in a proposed class action (U.S. District Court, Northern District of Illinois, Case No. 1:25-cv-02530), saying they believed they were talking to creators and were talking to agency chatters instead. However that case ends, chat authenticity has become a liability question — and the account it plays out on carries your name.

What does my contract actually let me do?

Read it for four things: notice period, exit fee, what happens to income already in flight, and who owns the account and content. Reporting on this industry has found commissions around 50%, five-figure exit demands, and mostly unlicensed operators, so verify what your paperwork actually says, not what you remember agreeing to.

Day four is the full read, end to end, even if you think you know it. Alongside the four levers above, check the base your share is computed on. In the Unruly case the agency charged 25 percent of all monthly gross revenue (gross, meaning before OnlyFans’ own 20 percent comes off), and the gap between a cut of gross and a cut of net compounds every month you leave it unexamined.

The wider market for these contracts is rough. IBTimes UK, reporting on a BBC investigation into OnlyFans management agencies, describes commissions commonly around 50 percent and sometimes as high as 70, managers holding creators’ logins and payment tools, one creator told she would have to pay £10,000 to leave, and an industry operating largely without licenses. None of that means your agency does any of it. It means the range of what paperwork in this industry says is wide enough that yours deserves an actual reading.

It is also on record that these contracts do not always survive scrutiny. In a Determination of Controversy dated August 20, 2024, the California Labor Commissioner declared the management contract in Stage v. Unruly Agency LLC illegal, void and unenforceable, finding that running the creator’s daily fan chats and negotiating prices without a talent agency license crossed into licensed territory under California law; a parallel civil case ran in Los Angeles County Superior Court (Case No. 22STCV06689). That is one state’s law applied to one contract’s facts — it tells you nothing certain about yours, and this is not legal advice. It does tell you that “the contract says so” is the start of a conversation, not the end of one.

For the clause-by-clause read (termination, exclusivity, post-termination shares), use the contracts guide, and pair it with who owns your account and content if the paperwork goes quiet on ownership.

How do I run the performance conversation so it produces something usable?

On a call you scheduled, with your numbers in front of you, and in writing afterwards. Present the gap, ask what their plan is, and set a deadline: a written plan within 48 hours, first measurable movement inside 30 days. An agency that cannot produce a plan in two days does not have one.

This is not the informal check-in from the signs guide — you are past that. Day five is a scheduled call with an agenda you set, and it exists to produce a document. Bring three numbers and one question. The numbers: your revenue trend across the three windows you pulled, message income specifically, and output — what actually shipped. The question: what is your plan to turn this around? Then stop talking. A team with a plan answers with specifics and dates. A team without one answers with the market.

On that point, the market excuse deserves numbers of its own. In OnlyFans’ 2024 financial year, fan payments hit $7.2 billion, creator accounts grew 13 percent to 4,634,000 and fan accounts grew 24 percent to 377,456,000. Competition is real — those creator numbers cut both ways — but a platform growing on both sides is a weak villain for a flat six months.

Before the call ends, fix the follow-up: a written plan within 48 hours (what changes, who owns each change, which number it should move and by when) and first measurable movement inside 30 days. That timeline is the standard from the signs guide: judge the process from week one and the results from day 30, and what should never take a month is someone getting back to you. Confirm all of it in an email the same day. That email is the artifact the rest of the week hangs on.

When should I skip the talk and start the exit today?

When the problem is control, not performance. If you cannot see your own dashboard, if payouts route through accounts that are not yours, if leaving carries a fee nobody can show you in writing, or if nobody answers you for weeks, negotiation has no counterparty. Start the exit and document as you go.

Performance problems get conversations; control problems get exits. The BBC-sourced reporting above sketches the shape: managers holding logins and payment tools, five-figure demands to leave, no license anywhere in the chain. If that is your situation, the seven-day clock is a courtesy the situation has not earned.

Silence is the other trigger. A question about your own money that goes unanswered for weeks is an answer. You are not obliged to schedule a performance conversation with a counterparty who does not reply; you are only obliged to document that you asked.

The mechanics are already written up, so this guide will not repeat them: the leave guide has the exit sequence (notice, backups, the order you take your logins back) and the lock-out guide covers what to do if access starts disappearing while you move. If your contract contains something that reads like it forbids all of this, remember that at least one such contract, put in front of a regulator, was declared void. Then take yours to a lawyer who reads contracts for a living. A few hundred dollars against a five-figure exit demand is not a hard trade.

Where should I stand by day seven — stay, renegotiate, or leave?

With one of three positions, in writing, even if only to yourself. Stay because the plan is credible and dated. Renegotiate because the work is fine but the terms are not. Leave because the plan never arrived. A decision made by exhaustion is the only wrong outcome of the week.

Stay is the right answer when the 48-hour plan landed on time, reads specific, and the person who sent it could answer the who-chats-my-account question without flinching. Staying is not dropping it: the plan gets a 30-day checkpoint in your calendar, with the numbers you will judge it by written down today, not remembered later.

Renegotiate when the work is defensible but the terms are not: a share computed on gross, chat staffing you had to discover yourself, reporting that only exists when you chase it. You now hold a file most creators never assemble, which is the entire reason this week started with collection instead of a phone call.

Leave when the plan never arrived, arrived without dates, or arrived and you realized you no longer believe them. Be honest about the cost: switching dips before it climbs, because the incoming team has to learn your fan base before pricing and scripts get rebuilt. The switch guide walks that handover and says so plainly. A shallow dip you planned for beats a flat quarter you kept extending. And if you have been burned before, the vetting guide is the checklist for whoever comes next.

What if the week shows the agency isn’t the problem?

Then the week was worth it. Sometimes the file shows a ceiling that has nothing to do with the agency: a traffic source that dried up, pricing that stopped matching your audience, a niche that moved. Firing a competent team to escape a problem that travels with you buys the dip without the fix.

Run the counterfactual before you fire anyone. If output held steady, the chat threads read fine, and revenue slid anyway, the cause may sit upstream of the agency. The income-drop guide walks that diagnosis cause by cause, and it is a better use of day seven than a termination letter aimed at the wrong target.

The same goes for ceilings. If you have been parked at the same number for two quarters while the agency competently maintains it, the question is not who to blame but what would break the plateau — a strategy problem, and a different article: why your revenue is stuck at $20K.

Either way, the week bought you something you did not have when it started: a documented picture of what your agency actually does all day, and a decision you can defend to yourself in three months. That holds whichever of the three doors you take.