Productized clip tiers and where they break
The 20 / 40 / 80 clips-per-month agency ladder mapped onto solo capacity, and why client review cycles rather than render time set the real ceiling.
Agencies selling AI-produced UGC have converged on a three-rung ladder, and one of the clearest public examples comes from Admiral Media: roughly 20 clips a month at up to 15 seconds for €4,000, 40 clips at up to 30 seconds for €9,600, and 80 clips at up to 60 seconds for €21,500. The ladder is real and the pricing is published. What is not published is headcount, and that omission is the entire problem for anyone reading those tiers and wondering which one they could sell solo.
The intuitive answer is that the ceiling is render throughput. It isn't. Render throughput stopped being the binding constraint the moment generation dropped to minutes, and what replaced it is the part nobody productizes: the number of times a human outside your business has to look at something and say yes.
The ladder, and what each rung implies
| Tier | Clip count | Max length | Published price |
|---|---|---|---|
| Entry | ~20/month | 15s | €4,000/mo |
| Mid | ~40/month | 30s | €9,600/mo |
| Top | ~80/month | 60s | €21,500/mo |
Two things are worth noticing before capacity enters the picture. The clip count exactly doubles between rungs, but the price more than doubles both times, so per-clip revenue climbs instead of falling — roughly €200, €240 and €269 a clip across the three tiers. Nobody is discounting for volume here. What the higher rungs charge for is a longer maximum duration and, implicitly, a wider creative remit. And that length ceiling changes the production shape far more than the count does. A 15-second clip is a hook and a payoff. A 60-second clip is a script, and a script needs approval before anything renders.
What actually consumes the hours
Take one finished clip through the whole loop and time each stage honestly. The numbers below are the shape, not a benchmark — your own will differ, and the point is the ratio between the rows, not the values.
| Stage | Who does it | Elapsed time | Your time |
|---|---|---|---|
| Brief interpretation | You | Minutes | Minutes |
| Prompt and generate | You | Minutes | Minutes |
| Select the take | You | Minutes | Minutes |
| Edit, caption, format | You | Minutes to an hour | Same |
| Client review round 1 | Client | 1–5 days | Zero |
| Revision | You | Minutes | Minutes |
| Client review round 2 | Client | 1–5 days | Zero |
| Publish or hand off | Either | Minutes | Minutes |
The two bolded rows consume no working time at all and dominate the calendar completely. This is the thing that breaks naive capacity math: you can produce 80 clips in a fortnight and still miss the month, because 80 clips means somewhere between 80 and 160 discrete approval events at a client whose marketing lead reviews things on Thursdays.
The credit side of this is well behaved by comparison. Iterating in 480p preview costs nothing — the free pass carries a short per-user cooldown, and the single charge lands on the final export no matter how many clips are on the timeline. So a five-round revision cycle and a one-round revision cycle cost the same to produce. They do not cost the same to live through.
Where each rung breaks for a solo operator
20 clips a month is comfortably solo. That's roughly one clip per working day, with slack. The review load lands at 20–40 approval events spread across four weeks, which one client contact can absorb without becoming a bottleneck. At this volume the work stays production-bound, which is the healthy state: the calendar is set by your own throughput rather than by someone else's inbox.
40 clips a month is where review breaks first. Not your capacity — theirs. Forty clips is 40 to 80 approval events, and a single marketing manager with a day job cannot clear that volume in weekly batches without falling behind. The failure mode is specific and recognisable: a queue of finished-but-unapproved clips builds, the month ends with 31 published, and the client's honest impression is that you underdelivered. You didn't. They did, and the contract said nothing about it.
80 clips a month is a different business, not a bigger version of the same one. At that count the client cannot be the reviewer in any per-clip sense. Either the contract moves to standing approval against a pre-agreed brief with spot-checks, or someone on your side is a full-time producer whose job is chasing sign-off. There is no third option. Note also that the step from the mid rung to the top adds roughly €12,000 a month of revenue, which is the order of magnitude a second full-time hire costs to carry. Read the top rung as a two-person tier and the ladder stops looking like a bigger version of the entry one.
Worth naming plainly: there is no public data on how many solo operators actually sustain these retainers. The rate cards are real; the fill rates are not published. Treat the ladder as a menu, not as evidence that the top rung is reachable alone.
The three variables that set your real ceiling
Forget clip counts for a moment and measure these three instead. Multiply them and you have your number.
Review rounds per clip. If your average is 1.2, you're near the floor and volume scales. If it's 2.5, your effective capacity is roughly half what your production speed suggests. This is the highest-leverage number in the whole business and almost nobody tracks it.
Approvers on the client side. One named person is workable. Two is a negotiation between them that you host. Three or more and the clip is not being reviewed, it is being litigated, and no clip count is safe.
Reroll rate. Not for the credits — though budgeting against the shots you throw away is worth doing — but because a high reroll rate means the brief is underspecified, and an underspecified brief produces more review rounds downstream. Rerolls and review rounds are the same disease showing up at two different stages.
Contract terms that raise the ceiling
Every one of these buys capacity without buying a person.
- One named approver, in writing. Not "the marketing team." A name. If that person is out, an alternate is named in advance.
- Batched review windows. Clips are delivered in batches on fixed days and reviewed in one sitting. Ten clips reviewed together take less than half the client's attention of ten clips reviewed separately, and it removes the daily interruption that makes clients resent the retainer.
- A revision cap, plus what happens past it. Two rounds included, further rounds billed or deferred to next month's allocation. The cap is rarely invoked; its job is to make round three feel like a decision rather than a default.
- Deemed approval after N business days. The single clause that most reliably rescues a 40-clip tier. Unreviewed clips publish, and the client knows that in advance.
- A standing brief that survives the month. Agreed once, referenced thereafter. This is what turns "does this fit the brand" into a check rather than a conversation. A reusable saved workflow does the same job on your side, and a recurring series does it for the slots that genuinely repeat.
The review loop itself can be tightened independently of the contract. 480p previews plus no-login share links remove the two frictions that stretch a review round: waiting for a paid export before anyone can look, and asking a client to log into something before they can watch. Neither is glamorous and both compress calendar days, which is the only currency that matters at the mid tier.
Pricing the tier you can actually hold
The honest sequence is to fix the ceiling first and price into it, rather than sell the tier and discover the ceiling. Run one month at the rung below what you think you can do, measure review rounds per clip, then decide. Pricing an AI video service is far easier once you have your own review-round number, because it converts a guess about capacity into arithmetic.
The tier that ruins agencies is the one sold on production capability and delivered against review capacity. How agencies scale client video output is mostly the story of that gap closing, and the operators who survive the mid rung are the ones who wrote the approval mechanics into the contract before they needed them. If you're building the offer from scratch rather than upgrading one, the agency overview covers the surrounding pieces.
FAQ
If render time isn't the constraint, why does clip count still get sold?
Because it's the only unit a client can price. Nobody buys "two review rounds per clip with a three-day deemed-approval window," but everyone understands 40 clips. Clip count is the invoice line; review capacity is the delivery reality. The job is to keep them from diverging far enough to lose the account.
Should the length ceiling change my capacity estimate?
More than the count does. A 15-second clip can be briefed in a sentence and approved on a glance. A 60-second clip usually needs script approval before render, which inserts an entire extra review cycle upstream of production. When you move from a 30-second tier to a 60-second tier, assume you added a round per clip and check whether the price change covers it.
Can I run 80 clips a month solo if I automate hard enough?
Automation compresses your side of the loop, which is already the fast side. It does nothing to the client's side, which is the slow side. The only thing that makes 80 workable alone is a contract where most clips do not require individual sign-off — standing approval against a locked brief, spot-checked. If the client won't agree to that, the tier isn't solo-viable no matter how good the pipeline is.
What's the earliest signal that a tier is about to break?
Unapproved finished work accumulating. Track it as a count, weekly. If the number of clips sitting in "delivered, awaiting response" is rising two weeks running, the tier is past its ceiling and the fix is contractual rather than operational. Waiting for it to resolve itself is how the month ends with a delivery shortfall you'll be blamed for.