Oct 29, 2026Material & Compliance
A Quality Scoring Rubric for Incoming Tableware
"Looks fine" is not an inspection. The scoring rubric — dimensions, weight, glaze, sound, edge — that turns incoming batches into numbers, verdicts and accountability.
"Looks fine" is not an inspection. The incoming batch that gets a glance gets drift, defects and disputes; the batch that gets scored gets numbers, verdicts and accountability. The quality scoring rubric is the inspection's instrument: the criteria, the scales, the thresholds — turning the eyeball into a system. This guide builds the rubric.
Why Score Instead of Look
The glance fails three ways: it adapts (the eye accepts the batch after ten minutes of looking), it forgets (last month's batch is a memory, not a reference), and it argues (the drift debate without numbers is opinion against opinion). Scoring fixes all three: the criteria fixed, the scale fixed, the reference fixed — the batch's verdict is a number against a standard. The rubric's promise: the inspection that two different inspectors would score the same.
The Rubric's Criteria
The scoring lines, per piece class:
Criterion | What it reads | Method |
|---|---|---|
Dimensions | Diameter, depth, capacity vs spec | Calipers, the capacity pour |
Weight | The mass vs the sealed reference | The scale — the hand's honest instrument |
Glaze color | The tone vs the sealed reference | Daylight or light box, against the standard |
Glaze surface | Defects: pinholes, blisters, crazing | Raking light across the face |
Edge and rim | Chips, rough spots, the profile's fidelity | The gloved finger's pass, the light test |
Decoration | Registration, color match, coverage | Against the golden standard's decoration |
Sound | The ring (the body's integrity) | The tap test — the dull note is the crack |
Stack behavior | Nesting depth, bridging | The actual stack, counted |
Foot and ring | Flatness, roughness (the table's scratch risk) | The flat-surface spin test |
The criteria's discipline: each scored on a defined scale (the 0-2 or 0-5 per line), against the sealed reference, at the defined light. The rubric is a form, not a feeling.
The Scales and Thresholds
The scoring's arithmetic: per criterion, the scale (0 = fail, 1 = marginal, 2 = pass — or the five-point where the stakes deserve it), the weighting (the color drift that matters to the brand weighs more than the capacity's milliliters), and the batch threshold (the acceptance level: the AQL logic where the sampling plan sets how many pieces are checked and what passes). The thresholds' discipline: set before the batch arrives (the contract's quality annex — the thresholds are the supplier's too), and applied without negotiation at the dock. The scale's honesty: the marginal piece is scored marginal — the 1 that gets rounded to a 2 is the drift's first lie.
The Sampling Plan
You cannot score everything; you sample correctly: the sampling plan — how many pieces per batch (the AQL-style tables per lot size), the draw's randomness (from the stack's depth, not the top), the per-criterion sampling (the weight on more pieces, the destructive tests on fewer). The plan's discipline: stated in the contract (the supplier knows the inspection's depth), executed the same way every time (the plan's consistency is the scores' comparability). The sampling plan is the rubric's statistics: the batch's verdict from a defensible sample.
The Verdict Tree
The scores land; the tree decides: ① pass (the batch enters service, the ledger records), ② marginal pass (the concession documented — the drift within tolerance, priced or noted), ③ fail (the remedy per the contract: reject, replace, rework — and the freight's ownership per the terms), ④ the pattern read (the failing criteria across batches — the supplier's systemic issue opens the corrective-action conversation). The tree's discipline: the verdicts recorded per batch (the ledger that builds the supplier's track record), and the marginal's honesty (the concession that becomes a habit is a threshold in disguise).
The Reference Infrastructure
The rubric runs on references: the sealed golden standard per piece class (the physical truth), the calibrated instruments (the scale, the calipers, the light box), the daylight or standard lamp (the color's honest stage), and the documentation (the spec sheet per class, the thresholds per criterion). The infrastructure's discipline: maintained, calibrated, and versioned with the spec. The rubric without references is a form without a ruler.
The Training Effect
The rubric trains the inspectors: the calibration sessions (two inspectors, one batch, the scores compared — the gaps discussed), the reference drills (the marginal pieces that teach the line between 1 and 2), and the rubric's living review (the criteria that never discriminate get revised). The training's payoff: the inspection that survives personnel changes — the scores comparable across years and hands. The rubric is also a teaching document: the fleet's quality culture written as criteria.
The Supplier Feedback Loop
The scores travel back: the batch report to the supplier (the scored rubric, the photos, the verdict), the trend line shared (the supplier's score history — the improvement or the pattern), and the corrective actions tracked (the root causes, the verifications). The loop's discipline: factual and regular (the supplier that sees its scores manages them), and tied to the commercial relationship (the score history is the renewal conversation's data). The feedback loop turns inspection into quality partnership — or into the exit case, documented.
Frequently Asked Questions
Why score instead of look? The glance adapts, forgets and argues; the rubric fixes criteria, scale and reference — the verdict two inspectors would score the same.
What does the rubric read? Dimensions, weight, glaze color and surface, edges, decoration, sound, stack behavior, foot flatness — each scored on a defined scale against the sealed reference.
How is the sampling plan set? AQL-style tables per lot size, drawn randomly from the stack's depth, stated in the contract. The plan's consistency is the scores' comparability.
What happens to a marginal batch? The documented concession: the drift within tolerance, priced or noted — and watched, because the concession that becomes habit is a threshold in disguise.
Why share scores with the supplier? The factual loop: the supplier that sees its scores manages them; the score history is the renewal conversation's data — or the exit case, documented.
