Field notes / 2026-11-12
Five things to check on the shop floor — before you ask a single question
The shop floor does not lie; the person hosting you might. Before pricing a precision-parts supplier, spend twenty minutes looking at five things. That alone is seventy percent of the judgement — everything after is verification.
2026-11-12 · 6 min read
Why looking beats asking
A quotation can be written by someone else. Sales talk can be rehearsed. Shop-floor condition cannot be faked for a visit, because it is the residue of daily production habits — nobody can clean it up overnight.
There is a more practical reason too. Asking gets you the version the supplier chose to give; looking gets you the version he did not plan to show. The field anchors on this site — thirteen fields per company — are built exactly for this: what you see on site goes straight into the matching field afterwards.
First: machine utilisation
How many machines are actually running? Lights on but spindles idle means no orders and thin cash flow. The opposite — fully booked — means your small order will be queued last.
This maps to the capacity field in our records. We only accept in-production figures; planned capacity and actual output are strictly separated. Public sources almost never contain this — which is precisely why it needs site verification.
changying-precision · Monthly capacity → #f-capacity
Second: in-house inspection capability
Look for a CMM, a vision measuring machine, a surface roughness tester. If none of them are there and the shop still accepts ±0.01 work, they are relying on calipers and luck.
This maps to the inspection field. The standard is not "has equipment" but "model + unit count + calibration still valid" — all three stated, or the field stays Pending.
changying-precision · Inspection capability → #f-inspection
Third: calibration labels on hand tools
Calipers and micrometers carry calibration stickers. Read the dates. Labels expired by more than three months mean management exists only on the wall.
This is the cheapest item of the five: you do not need to go deep into the shop. The tools at the operator's bench are enough.
Fourth: traveller cards on work in progress
Does every pallet and every material basket carry a process card — which machine, which operation, which operator? If yes, defects can be traced to a step. If no, every rework becomes an argument.
For thin-wall housings and joint shells — parts where consistency matters — the deciding factor is never one good machine. It is traceability. A shop without traveller cards can make a good sample; it will drift in production.
changying-precision · Processes and scope → #f-scope
Fifth: the floor and the tooling area
Aluminium chips piled into corners, tool holders mixed in one drawer — such a shop can handle one-off parts, but consistency will drift.
A shop with tools sorted by specification and labelled material zones is not automatically a good shop. But the reverse is certain: fail all five, and no batch order should go there.
changying-precision · Key equipment → #f-equipment
What to conclude after the walk
Four of five: worth discussing a long-term relationship. Two or fewer: one-off orders only, payment on delivery only.
After looking, sit down and close only three gaps: key equipment models, how inspection is guaranteed, and which day the lead time starts counting. The answers go straight into the matching fields of the company record — which is also why fields on this site are left empty. An empty field is not missing data; it means nobody has verified it on site yet. Whoever goes first holds the complete version of the table.