CMF is engineering: materials that survive year three.

Every product looks good in launch photography. The honest test happens around year three: after a thousand wipe-downs with whatever solvent the site actually uses, after UV, after keys in the same bag, after the texture has met ten thousand gloved grips. Products age in ways mood boards never predict — and whether they age with dignity is decided early, in choices that look cosmetic.

The decoration trap

The trap is sequencing. When CMF arrives at the end — "make it look premium" — the material palette is already locked by tooling decisions made without it, and the only lever left is paint. Painted finishes on hard-use products are a maintenance contract you've signed on the user's behalf. The alternative is to treat colour, material and finish as engineering inputs from the first architecture sketch, chosen with the same rigour as a bearing or a seal.

Questions a finish must answer

Every surface specification we sign answers a checklist that has nothing to do with taste: What chemistry will clean it, and how many cycles does the supplier's data actually cover? How does the texture wear at contact points, and does worn look broken or look patinated? Can the tool actually deliver that gloss level at that draft angle, or are we specifying rejects? What does it cost per part at your volume — and at half your volume, if the forecast is wrong? How does it photograph, since the render is making promises the moulding must keep?

Specify it so production can hold it

A CMF direction that lives in a slide is a suggestion. A CMF system that survives production names materials and grades, texture references, gloss ranges with tolerances, and approved suppliers where it matters. That document is the difference between batch one and batch nine matching. It's also brand infrastructure: the finish language is part of what makes a product family recognisably yours, which is why our industrial design practice treats CMF and design language as one system, resolved with engineering against the processes you'll actually run.

Premium isn't a look. It's the absence of regret at year three.