There are two ways to price creative production, and the difference between them decides how the whole project feels.
A day rate says: here is what an artist costs per day, here is roughly how many days we think it will take. It is easy to produce, easy to compare between vendors, and it tells you almost nothing about your actual project.
A shot count says: we watched your cut or read your asset list, here is every item, sorted by complexity, with a number attached to each. It takes us longer to produce. It is the only one of the two that survives contact with reality.
Why the difference matters
Halfway through every project, something changes. A sequence gets recut. A character is cut and two are added. The publisher wants the level in snow as well as rain.
With a day rate, you now have a negotiation. How many extra days? Who decides? Was that in the original estimate? Nobody can answer from the document, so the answer comes from whoever is more stubborn.
With a shot count, you have arithmetic. Six shots came out, four went in, three of the new ones are band B and one is band C. The new number is not a matter of opinion — it is the same rates applied to a different list, and it takes ten minutes rather than a week and a difficult call.
That is the entire argument. A shot count converts the most stressful conversation in production into a calculation.
What a bid actually looks like
For a VFX or animation job we send a spreadsheet with a row per shot:
- Shot ID and duration
- What the work is, in one line
- A complexity band — A, B or C, defined explicitly at the top of the sheet
- Assumptions specific to that shot ("assumes clean plate supplied")
- Days, and therefore cost
For game art, the same structure with assets instead of shots: asset name, band, polygon and texture budget, whether concept is included.
The assumptions column is the important one and the one most bids leave out. Half of all overruns trace back to something the vendor quietly assumed and the client quietly assumed differently. Writing them down does not prevent the disagreement — it just moves it to week one, where it is free.
The bands are the honest part
Band definitions are where a bid is either useful or decorative. Ours read like this, adjusted per project:
- Band A — the work is well understood and similar to something on the list already. Low risk. If we are wrong, we are wrong by hours.
- Band B — a known technique applied to a new problem. Some iteration expected. If we are wrong, we are wrong by a day.
- Band C — we have to solve something before we can build it. Genuinely uncertain. We will say what would reduce the uncertainty, and often it is a half-day test we would rather do before the project starts than during it.
If everything in a bid is band A, the vendor has not read the brief carefully. Real projects always have a few Cs, and a bid that hides them is not cheaper — it is just wrong later.
What we do about being wrong
We will be wrong on some shots. Everyone is. The question is who absorbs it.
Our rule: if we underestimated, that is ours. A shot we banded A that turns out to be B is our misjudgement, and we eat it. If the brief changed — new reference, new note, new requirement that was not on the sheet — that is a change, and we price it as one, before doing the work rather than after.
The distinction is not always perfectly clean, and when it is genuinely ambiguous we would rather split it than argue about it. But having the rule written down means the conversation starts from a shared principle instead of from whoever emailed first.
A quote is a prediction. The useful part is not the number — it is being able to see how the number was built, so you know what happens to it when something moves.
What we need to build one
Enough to be specific: the cut with timecode, or the asset list; the delivery spec; the date; and whatever reference exists for the look. If those do not exist yet, we will give you a range with the assumptions stated, and replace it with a real count as soon as they do.
What we will not do is send a single number with no working. It looks decisive, and it is the least informative thing a studio can hand you.
Comments
Loading comments…