There is a structural tension in every hand-finished collectible: the process is manual, and the buyer expects consistency. Those two facts do not reconcile themselves. They are reconciled by an agreement made before production — a written acceptance standard — and by nothing else.
When that agreement does not exist, the argument at the end is not about quality. It is about expectations, and expectations are unwinnable arguments.
What an acceptance standard is
An acceptance standard is a short document, illustrated with photographs, that does three things:
- Defines defect classes — which deviations are unacceptable, which are tolerable, and which are simply part of a hand-applied finish.
- Names the reference — the physical sample that everything is compared against.
- Fixes the inspection method — viewing distance, lighting, quantity inspected, and what happens when a result is contested.
It is written before production starts. A standard written after the first batch is a negotiation, not a standard.
The three classes, and where the line actually falls
Every workable standard we have seen uses three classes. The naming varies; the logic does not.
| Class | What belongs here | Treatment |
|---|---|---|
| Critical | Structural failure, sharp edges or points, loose or detachable parts, finish that flakes or lifts, contamination in the surface | Rejected. Not a matter of degree |
| Major | Visible runs or sags, misplaced or missing paint on a focal area, colour outside the agreed reference, visible seam where it should be hidden | Rejected on the unit; a rate triggers review of the batch |
| Minor | Slight variation in coverage on a low-visibility surface, faint brush texture, small deviations only visible at close range | Accepted within a stated limit |
The hard part is not the table. It is the boundary between major and minor, and that boundary has to be photographed, not described. Written descriptions of colour and finish break down immediately when two people look at the same piece and disagree.
The single most useful rule: any defect that is visible at normal display distance is not minor. Buyers view collectibles from where they will be displayed, not through a loupe, and the standard should reflect that.
The defect class that gets argued about at the wrong stage
There is one category that reliably causes a wasted dispute, because it is inspected at the wrong point in the line. Air trapped against a tool surface leaves small voids in the casting itself — pinholes and surface porosity — and no volume of paint removes them. At a finish inspection they present as a paint fault, and the conversation starts in the wrong place.
Three things follow, and the acceptance standard should state them:
- They belong in the cast inspection, not the paint inspection. Once the piece is painted, the cause is no longer visible and the argument becomes about coverage rather than about the casting.
- Porosity clusters rather than spreading evenly. It is normally a small number of units rather than a low background rate across the batch, which makes it a batch signal — a rate, not a per-unit judgement.
- Raking light finds them; diffuse light hides them. If the standard fixes one diffuse viewing condition and nothing else, this class will be under-detected in the factory and discovered by the collector instead.
That last point generalises. Most surface defects — porosity, sanding marks, faint sink marks, a slightly proud seam — are found by light across the surface rather than light onto it. A standard that names a viewing distance but not a light direction is only half specified.
The approved sample is the contract
Specifications describe. A sample decides. Where the two disagree, production follows the sample — which is exactly why the sample has to be approved formally.
Formal approval means a signed record referencing a specific physical object, not a message saying it looks good. Informal approval is the most common cause of a disagreement we cannot resolve, because both sides are remembering a different piece.
Keep at least two approved samples: one held by production, one held by you. A single reference sample that lives only at the factory is not a reference — it is an internal opinion.
Measuring colour without arguing about it
Colour is where hand-applied finishes and formal standards collide hardest, and it can be made manageable:
- Fix the lighting. Colour judgements under mixed light are meaningless. Agree on a defined viewing condition and use the same one at every inspection.
- Compare to a physical reference, not a screen. A rendered concept is a design intent, not a production target.
- Distinguish spot colour from gradient. Flat areas can be held tightly. Soft transitions and metallic sheens have to be allowed a wider window, and both sides should say so in advance.
- Agree where colour is critical. Sheet metal on a focal point is critical. The underside of a base is not. Making that explicit prevents most colour disputes before they start.
Inspecting, not glancing
A standard that says "inspected" without saying how is not a standard. Fix at minimum: the fraction of the batch inspected, the lighting, the viewing distance, and whether the piece is judged before or after packaging.
One point that buyers routinely underestimate: inspecting after packaging is a different activity with a different defect profile. Damage introduced by packing is invisible to a pre-pack inspection, and packing damage is what the collector actually receives.
What goes wrong most often
In order of frequency:
- A hero sample is approved, and no written standard follows. The sample sets an unachievable bar that nobody documented, and the first production batch is judged against a memory.
- The standard is written after the first batch. Now it is a compromise, and both sides know it.
- Minor defects are defined by size rather than visibility. A small mark in the wrong place matters more than a larger one in a hidden one.
- No agreed handling of a failed batch. The standard should state what happens next, not only what counts as failure.
- Cost pressure applied to inspection rather than to design. Reducing inspection on a hand-finished product moves defects from the factory to the customer, where they cost far more.
A short checklist to agree before production
- the physical reference sample, approved in writing, with a copy held by each side
- the three-class table, with photographs for the major/minor boundary
- the lighting and viewing condition used for inspection, including light direction — raking light for surface defects, diffuse light for colour
- inspection point: before packing, after packing, or both
- inspection quantity, and the result that triggers a batch review
- the handling procedure when a batch fails
- colour tolerance, split by critical and non-critical areas
None of this is legal language, and it does not need to be. It needs to be specific enough that two people looking at the same piece reach the same conclusion.
What we do not publish: the specifics of our process engineering. Production photography on this site is intentionally obscured. The standard described above is the part that belongs in your hands.
