A review meeting early last year: the render on the wall, every hand in the room up, the client signing without a single note. Eight weeks on, the CNC sample came back and the highlight running along the waist had a kink in it that nobody could unsee. Nobody had approved that kink. Nobody had been shown it either.
That is the pattern. The most expensive rework in appearance design almost always happens after the render review passes unanimously. Go back through the files and it lands on the same three invisible items: highlight flow the render draws but the part cannot hold, a parting line placed where it fights the form, grip dimensions left unverified until there is something to hold. All three are invisible in a render, which is exactly why the room stays quiet and the client nods so easily.
The Cost Jumps by Stage
Rework is not priced linearly. The same decision costs one thing at the render stage and another thing entirely once steel exists.
| Stage | What one form change takes | What it actually costs |
|---|---|---|
| Render stage | Done the same day | Designer overtime; almost no external cost |
| CNC sample stage | One to two weeks | Remake the sample, a few thousand RMB, and rebook the review |
| After the tool is cut | A month or more | Tool modification or a fresh tool, tens of thousands of RMB and up, plus a launch date that moves |
Read the middle row against the bottom one. Same sketch, same conversation, a jump of an order of magnitude in money and roughly four times the calendar. The point is not to get better at late fixes. It is to draw those three invisible items on the drawing, and get a signature against them, while the cheapest row still applies.
The Render You Sign Off Is the Least Trustworthy Thing in the Project
The image in the sign-off deck is the most polished lie in the whole programme. Environment light picked from the best-looking HDRI. A rim highlight brushed in during post. Reflectance parked at the value that flatters the form most. The machine the client sees is at least 20% better looking than the one that comes off the machine shop floor, and no amount of good faith on either side changes that.
The consequence is behavioural, not technical. When the part arrives and does not match the picture, the client's first sentence is not "something here is slightly off." It is "the form is wrong, do another version." If the designer cannot point at the exact millimetre where the two diverge, the project slides from tuning parameters to redoing the form — a completely different order of magnitude, and the one that eats the schedule.
So the review meeting does not show a render on its own. The same form version goes up with three attachments:
- The highlight flow map — where the band runs, where it breaks, how it terminates.
- The parting line map — the line on the form, in position, with the draft direction marked.
- Cross-section dimensions at the grip points — the actual numbers the hand will meet.
The client signs those three. The pretty picture is what they take home to show their boss.
Highlight Flow Is Drawn, Not Built
This is the first source of appearance rework and the quietest one. In a render, you want a bright edge to run along the waistline: add a highlight map, five seconds, done. In the part, that same edge has to be earned by curvature. Class A surfacing, neighbouring faces at G2 continuity, curvature variation smooth enough that the eye reads one continuous line.
A case from our files. The form stage built a transition out of several patches. The render looked crisp and the review liked it. When it went to Class A, the curvature change at the seam was too abrupt, and the highlight stuttered exactly there — a thin S shape, a few millimetres long, brighter than any parting line in the room under real light. On the sample it read as a defect. On the drawing it had not existed.
The fix is scheduling, and it is free. Check with zebra stripes and curvature combs at the form stage, not after the Class A surface is finished. Decide at sketch level where the highlight starts, where it ends, where it breaks, and how the break is terminated. Two rules we hold to: a highlight must not die on a fillet, and it should not run straight through the parting line. Those are the two places where it shatters, and both are visible from a metre away once the part is gloss.
There is a cheaper move than more surfacing hours: increase the radius. A softer transition holds a steadier highlight. The cost is that the form goes soft and loses its edge. That is a trade, and the designer makes it — not the surfacing software, which will happily produce a continuous surface with no character at all.
Hand Feel Is Decided the Moment You Hold the Sample
Half the information in a handheld product lives in the hand, and none of it survives a render. Palm width commonly runs 75–90 mm. The comfortable diameter for a gripped cylinder mostly lands somewhere in the 30s to 40s of millimetres. The numbers are easy to look up. Landing them on this particular form while keeping the weight distribution sensible is the actual work.
Find out at the CNC sample that it cannot be gripped and you are no longer changing numbers. You are changing the form: body narrows, buttons move, the internal stack is re-laid out, and every render hour spent so far is written off. So grip validation moves earlier. Before the render is frozen, we cut a clay or 3D-printed rough model — no surface finish, no paint, just a shape you can close your hand around and pass to three people who have never seen the project. One day of that blocks two weeks of rework later.
Parting Lines and CMF Are Design Decisions
A parting line is one millimetre wide on a drawing, which is why it gets filed under tooling. It is a visual dividing line, and put in the wrong place it scatters the whole character of the machine.
Hand the form to a tool shop and they ask one question: can this line run straight? They will choose the easiest, most release-friendly position, which is usually the widest contour of the part — very often the ugliest available spot, slicing across the fullest part of the surface. Appearance faces normally need 0.5 to 1 degree of draft, and that angle nudges the visual run of the surface, most obviously on dark gloss parts. Parting line position and draft direction get fixed inside the form, and marked on the release drawing. If you do not own that decision, somebody else takes it for you.
CMF behaves the same way. The same form in black gloss PC and in fine matte dark grey demands a whole grade apart in surface quality. Black gloss enlarges everything: sink, weld lines, highlight stutter. Matte swallows most of it. Set the colour and texture last and the Class A standard you built for gloss black was spent for nothing.
We hold one internal line: before the appearance sign-off review, CMF must be fixed at least to colour family and surface finish grade. Below that, it is not signed off. Colour and texture are already part of how the form speaks; they are not a finishing task.
The Questions That Belong on the Drawing
Four items decide this: highlight flow, parting line, grip dimensions, CMF grade. If any one of them is pushed past the structural review to be dealt with later, the rework has already started — the only open question is which row of the table above you will be paying in.
Most projects do not fail on design ability. They move questions that belonged on the drawing to the sample stage, and then ask a physical object to answer them. It will. Just not cheaply.
Sources: Compiled from Hezi Industrial Design's own project experience; no external citations.
Article by Wang Ran / Hezi Industrial Design — Dongguan, China. Originally published at hezidesign.com.
Top comments (0)