Revenue dropped after you signed with an agency — walk the causes in order

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

The drop showed up after you signed, and the timing feels like the verdict. It isn’t, yet. “After” and “because” are different claims, and mixing them up is expensive in both directions: fire a competent team over a market slowdown and you pay the switching cost for nothing; sit politely through real damage and you pay for that instead. This is the agency fork of our general guide to income drops. That one covers every cause for every creator; this one assumes the drop coincides with signing and does one job: attribution. Six causes, walked in order, cheapest check first. By the end you should know which one is yours, and be able to show your working.

Are you looking at the same numbers you looked at before you signed?

Check the measurement before anything else. Before signing you probably watched gross fan payments in the OnlyFans dashboard; now you may be reading your payout after the platform’s 20% and after the agency’s share. Same page, three different numbers. Put both periods on the same line before you compare anything.

Three numbers describe the same page. Gross fan payments: what fans actually spent. Your net, after the platform keeps its 20%: OnlyFans passes 80% of fan payments through to creators and paid out $5.80 billion that way in fiscal 2024. And your share, after the agency’s percentage comes out of the net. Before signing, most creators quote the gross number, because that is what the dashboard shows. After signing, the number in front of you is usually the one on the agency’s statement, two cuts further down. A “drop” that is really a definition change is the fastest case to close, and it happens more often than anyone involved likes to admit.

Alignment errors do the rest. A 28-day month against a 31-day month is a ten percent swing before anything real has happened, and a comparison built from screenshots and memory inherits every rounding along the way. Rebuild both periods from the platform’s own statements page, same line, same dates, full calendar months. If the drop shrinks to noise, you’re done — and where exactly a fair agency takes its percentage is worth reading before your next statement anyway.

Is this the normal onboarding dip — and how long is normal?

A handover dip is normal in our experience: typically two to four weeks, and how deep it goes depends mostly on whether DM coverage stayed continuous through the switch. That is experience across our own onboardings, not a study. Inside week four, hold. Past week six with no recovery, keep walking this list.

We’ve written about that dip in the switch guide, including the week-by-week handover plan that keeps it shallow, so we won’t re-teach it here. The mechanism is mundane: your regulars notice new voices in the chat, new scripts need calibrating against your actual spenders, and scheduled content runs out before the new pipeline fills. None of that is misconduct. It’s friction, and it fades.

What month one should produce (and what it realistically can’t) is covered in our first-30-days guide, so set expectations there rather than against your best month of last year. The boundary matters more than the average: a dip that bottoms out in week two or three and starts turning is the pattern. A line still falling in week six is a different animal, and the rest of this guide is for it.

Did the whole market move, or just your page?

The platform is still growing (gross revenue reached $7.22 billion in fiscal 2024, up 9%), so “the platform is dying” does not explain your drop. But growth has slowed hard since 2021, and accounts are multiplying faster than money, so average revenue per account is under pressure platform-wide.

The platform’s own fiscal 2024 numbers, as reported by Yahoo Finance: $7.22 billion in gross revenue, up 9% year-over-year. Growth — but a different order of growth from the era many baselines were set in. Gross revenue rose 118% in 2021, then 16%, 19% and 9% in the three years after. If your reference month sits in a faster year, today’s comparison softens without anyone touching your page.

The more useful number for a single creator is the ratio of accounts to money. Creator accounts grew 13% to 4.634 million in fiscal 2024 and fan accounts grew 24% to 377.5 million, while revenue grew 9%; trade press read the same filing as user growth outpacing financial gains. More accounts sharing money that grows more slowly means the average account earns less each year, as a structural matter, with nobody at fault.

Two consequences for your diagnosis. First, the market’s signature is a slope, not a step: a platform-wide softening does not produce a cliff on your signing date. Second, seasonality: there is no citable statistic on how OnlyFans spending moves through the year, so ignore the folklore and solve it with method. Compare your April to last April, your own months year-over-year, never to the month before.

Did your traffic drop before your revenue did?

OnlyFans has weak native search and discovery, so revenue on the platform is downstream of external traffic. If a promo account was restricted or a funnel died around the time you signed, revenue falls with no agency involvement at all. Check clicks and new subscribers by source before you check anything on-page.

OnlyFans does little discovery for you. Trade press describes searchability and discovery as an unresolved platform limitation, a gap wide enough that third-party search engines exist to fill it. The practical translation: your revenue depends on whatever brings people in from outside, and a break anywhere in that funnel lands on your revenue line even when the agency’s on-platform work is unchanged.

The most common break is a promo account. Adult creators losing access to social accounts, and the direct hit to their business that follows, has been documented since at least 2018. If a platform restricted your reach or removed an account around the time you signed, that alone can explain the line.

Decompose it. New subscribers fell while spend per fan held: the problem is upstream, in traffic. Subscribers held while spend per fan fell: the problem is on-page (pricing, chat, content), which is where agency work lives. And watch for the coincidence trap: onboarding usually touches link infrastructure. New tracking links, a retired promo handle, a rewritten bio — any of these can break traffic at exactly the moment you signed, while being fixable in an afternoon rather than a reason to leave.

What does agency-caused damage actually look like?

Execution damage shows as spend per fan falling while traffic holds: slower DM replies, generic scripts, price changes you did not approve, mass messages at a frequency that burns spenders. Conduct damage is rarer and more serious: the kind of behavior that, in the worst cases, ends up alleged in court filings.

Execution first, because it is the common kind. Reply times stretch and spenders drift. Scripts read generic and per-fan spend sags. Subscription or PPV prices change without sign-off. Mass messages go out at a rate that trains fans to ignore them. The posting schedule quietly thins. Each of these has a number attached, and the full catalog of which numbers to pull — and what good looks like — is in our underperformance guide. This guide only needs the pattern: on-page metrics falling while traffic holds.

Conduct damage is the far end, and worth describing precisely because it reaches courtrooms. Elizabeth Machabeli & Jane Doe v. Unruly Agency LLC, Tara Niknejad, and Nicky Gathrite, filed in Los Angeles Superior Court as Case No. 21STCV41395, alleges employee misclassification and wage theft, and alleges that account managers deceptively impersonated models in fan chats; two further suits were filed against the same agency in 2021, Case Nos. 21STCV31028 and 21STCV26060. These are plaintiff-side allegations, not court findings. The chat allegation is the instructive one for this guide either way: fan chat is exactly where spend per fan is made, and when fans sense they’re talking to a script, they stop buying — whoever runs your chats.

Most drops are not this. But it marks the top of the scale, so you know what actually belongs there — and what is ordinary friction being read too darkly.

How do you prove which cause it is before you act?

Build one table: the eight weeks before signing and every week since, with gross fan payments, new subscribers by source, spend per fan, DM share of revenue, and posting cadence. The cause sits where the first break is. No break means measurement. A slope means market. Traffic breaks upstream; agency damage breaks on-page while traffic holds.

The table takes an evening. Weekly rows, revenue columns pulled from the platform’s statements rather than anyone’s summary, activity columns from your own posting history. Then read it for the first break, because each cause signs differently. Measurement error leaves no break once definitions match. The market shows a slope that predates your signature. The onboarding dip breaks at handover and turns within four weeks. A traffic cause breaks upstream first, spend per fan intact. Agency damage breaks on-page while traffic holds. One honest caveat: keep your own promo activity roughly level across the windows you compare, or the table measures your attendance, not their work.

Then ask your agency for their version of the same weeks. A team actually doing the work has these numbers and can produce them quickly. What arrives, how fast, and how it’s framed is evidence too. Which numbers to insist on, and how to raise a shortfall before you decide anything, is the underperformance guide’s territory; we won’t restate the conversation here.

What do you do once you know the cause?

Match the response to the cause. Measurement: fix your reporting and stand down. Onboarding dip: hold until week four. Market: reset the baseline, not the team. Traffic: rebuild the funnel, with them or without them. Agency execution: raise it with specifics before you quit. Agency conduct: start the exit.

Routing, briefly. If it was measurement, tell your agency what you found — a shared definition of “revenue” prevents the next false alarm. If it was the dip, re-run the table at week four and let the trend speak. If it was the market, judge the team on relative performance against your own year-ago months, not against a 2021 slope nobody can bring back. If it was traffic, fixing the funnel is joint work — though whether they spotted the break before you did tells you something about the monitoring you’re paying for.

If it points at the agency, order still matters: the conversation comes before the exit, and the underperformance guide linked above owns that script. If the conversation changes nothing, switching without repeating the dip is one path and the exit checklist is the other. If what you found was conduct rather than sloppiness, leave the conversation out entirely and read the recovery guide once you’re out. And remember that causes stack: a market slope, a mild dip and a definition change can add up to one scary line with nobody at fault. Fix measurement first, subtract the market, then judge what’s left.