A figure that sells as a 90 mm keychain and one that stands 240 mm on a retail shelf can share the same drawing and cost wildly different amounts, because what a buyer is actually paying for is not plastic — it is mould geometry, part count, paint passes and the tolerance somebody is willing to defend at final inspection. Same art. Different animal.
Last spring a licensee sent a 95 mm character with six colour areas, two of which appeared on both the front and the back of the head. Six masks, not four, because the same colour on two faces needs two openings unless you enjoy overspray across a cheek. That single detail moved sampling by roughly a third, and nobody had mentioned it anywhere in the brief at all!
Who pays for that discovery? Nobody, if it surfaces in week one and the drawing is still fluid; everybody, if it surfaces after tooling. I think most bad quotes are not greedy factories. They are confident guesses made against a picture and a hope.
Two extra masks. One extra week. That is how a launch date dies.
Part count is the first real number. A hollow body moulded in one piece with an injected face plate is two tools. Split the arms so they can be moulded in a second colour and you are at four; four tools means four trial shots and four separate places where a seam can drift out of register. Add a snap-fit joint and you have bought yourself a tolerance stack that has to be measured, not eyeballed.
Wall thickness is the second. A 2.8 mm wall at 95 mm holds its shape and still ships at a carton weight a courier will not punish you for. Drop to 2.2 mm and the figure dents in the bag after a week of handling, which is a cost you pay in returns rather than in tooling.
Then comes colour, and colour is where the arguments live. Plastisol takes pigment unevenly across a long run, pad printing holds a fine line but hates deep recesses, and a spray mask either seats properly on a curved surface or it leaks at the edge. Six masks need six colour checks per shift. Three masks need three. That is a labour line, not a creative decision.
The fourth confirmation is unglamorous: who signs. A sample approved by an assistant who has never seen the licensor style guide is not an approval, it is a delay with a signature on it. Our rule is blunt — a physical chip, a written spec, and one name on the sheet.
Fifth is the market. A figure sold to adults as a collectible and a figure handed to a six-year-old are different products with different test routes, and the bill of materials has to be frozen before any laboratory sees it. Swap a coating afterwards and you are re-testing, not substituting.
What we ask for, in the order that actually saves time: a front and side sheet, the finished size in millimetres, the quantity band for the first run, the market and age grade, the date you need the sample in hand, and the unit cost you are trying to hit. Six lines. Most briefs arrive with two.
Send those six to [email protected] and the reply is a process route, a part count and a sampling plan — not a catalogue and not a range so wide it means nothing. That is the only reason chibe puts part counts in a first reply: because the number that matters is decided there, not in the negotiation afterwards.
If you only remember one thing, remember this. A quote is a hypothesis about how the object will be made. Give the factory enough to test the hypothesis and you will recieve a number you can plan against. Give it a JPEG and you will get a number that survives exactly until the first sample lands.