The number would have closed our quarter at a figure we don't usually see. I said no anyway, on the second call, then again on the third. The brand pushed back twice. They walked. Six months later, the creator who would have earned the largest check of his year is still on my roster, shipping the videos he planned to ship.
This is the part of agency work nobody puts on a deck: the deals you didn't take.
What the brief actually wanted
On paper it was a strong campaign. Three integrations, four months, a category the creator already owned, a budget at the top of what I see for an audience his size. I wanted to like it.
The terms underneath were a different document.
The brand wanted final script approval, line by line, with revisions until they signed off. They wanted a posting cadence that compressed three integrations into his strongest content quarter, replacing two of the tentpole videos his audience waits for. They wanted exclusivity in a category he was already half-promised to elsewhere.
None of that is unusual on its own. Stacked together, it was a deal that turned a creator with a voice into a content vendor for a season. I've watched what happens to creators who sign those. The audience doesn't come back at the same volume.
The line I don't cross
I've reviewed enough brand-scripted integrations to know the shape before it happens. Retention falls off a cliff at the read. The comments shift from interest to suspicion. The next three uploads cool off, because the algorithm has been told something changed.
The brand sees one campaign in isolation. The channel pays for it across the next quarter.
I told them on the second call that we could deliver everything they wanted on outcomes: the goals, the must-mention points, an accuracy review on our side. We couldn't hand over final script approval. The creator writes his own integration, or the audience smells it. That's the mechanism the whole deal runs on.
Why "fine" is the trap
Their answer was that they'd run campaigns this way before and it had been fine. I believe them.
That's the problem. "Fine" is why it keeps happening. Fine means views landed in the right zone and a deck got approved internally. Fine doesn't measure the creator who stops returning the agency's calls six months later, or the channel that never gets its reach back. The cost is real. It shows up on a different page than the one the campaign gets graded on. So agencies keep signing.
What we lost. What we kept.
We lost $120K in agency revenue and a roster-level number on top. The creator lost the biggest single check in his pipeline that quarter. Both numbers were real.
We kept the creator. He shipped his two tentpoles, and both beat his channel average. He took two smaller deals from brands that respected the format. His next quarter beat the one before it.
The brand we walked from went to a competitor agency. I watched the integrations roll out. They underperformed the benchmarks I track for deals that size, and the creator they used has had quieter uploads since.
The Yes That Costs Later
- Big check, this quarter only
- Creator burned out by month four
- Audience trust eroded by scripted reads
- Algorithm penalty on next 3 uploads
- Creator's pipeline thinner by month six
The No That Pays Forward
- Lost revenue in the immediate quarter
- Creator's audience intact and growing
- Tentpole videos shipped on schedule
- Two smaller deals from brands that fit
- Roster strength compounding into next year
A yes can cost more than a no. It bills you later.
The question that matters
If you're a brand, ask whether your agency has a line it won't cross on your behalf, especially when you're the one asking it to cross.
An agency that says yes to every brief the way it says yes to every brand is depleting your creator partner, not protecting them. By the time you notice, the creator who used to convert for you isn't the creator you signed.
So ask the agency you work with: what is the last deal they walked away from on your behalf? If they can't name one, they haven't been protecting anyone. They've been processing.