Skip to content

Workmanship criteria in amazon fba inspection: part 2

RomeMovies
The short answer

Workmanship disputes are settled long before the inspector arrives: write down what a defect looks like, sort findings into critical, major, and minor classes, and judge the lot against agreed limits. This second part of our workmanship series covers appearance criteria and the classification discipline that keeps verdicts factual.

Key takeaways

  1. Define each workmanship criterion with a measurable threshold and a photo reference, so inspectors deliver verdicts instead of impressions.
  2. Sort every finding into critical, major, or minor classes before counting, because the class decides the consequence more than the raw count does.
  3. Draw the sample at random across all finished cartons, following acceptance sampling discipline, so the result represents the whole lot[1].
  4. Treat destination-market technical rules as severity inputs, consistent with a trade framework that recognizes legitimate regulatory objectives[2].
  5. Confirm the criteria with the factory in writing before production, so inspection day verifies rather than negotiates.

Most FBA quality arguments are not about whether a flaw exists but about whether it matters. A faint scuff inside a lid, a slightly uneven seam, a small color variation between batches: the factory calls each of these normal, the seller calls them defects, and without written criteria both sides are merely exchanging opinions. Workmanship criteria convert those opinions into testable statements with thresholds, photographs, and consequences.

This second installment of our workmanship series addresses appearance and classification: how to describe visual defects precisely, how to sort findings into severity classes, how sampling turns individual findings into a lot verdict, and how destination-market rules influence severity. Later parts of the series take up functional testing and packaging verification, so this part stays focused on what the eye, the hand, and the nose can judge on the packing table.

The intended reader is the seller who writes the purchase order, not the inspector who executes it. Each section gives purchase-order language that can be copied into specifications, explains why the wording matters at the inspection table, and notes the evidence the report should return. Criteria written this way protect both sides: the seller gets enforceable standards and the factory gets achievable targets.

Describe appearance defects in testable language

Vague criteria produce vague verdicts. A requirement that goods be free of obvious scratches fails the moment anyone asks what obvious means, at what distance, and under which light. Testable language replaces adjectives with conditions: no scratch longer than a stated length visible at arm's length under normal indoor light, no stain larger than a stated diameter on visible surfaces, no color difference exceeding the approved shade band when compared side by side with the reference sample.

Photography carries half of this burden. For every defect type the specification names, attach two photographs showing the boundary between acceptable and reject, taken from the inspection distance under neutral light. Borderline photographs teach faster than paragraphs, especially across language differences between the seller and the production team. Update the photo set whenever a dispute reveals an uncovered case, so the library grows with experience instead of repeating arguments.

Odor and feel need the same discipline even though they resist rulers. State the check procedure rather than the sensation: ventilate the sample unit for a stated time, smell at a stated distance, and compare against the approved sample kept sealed since approval. For texture, name the comparison method, such as running a fingertip across the seam and comparing with the reference piece. Procedures make sensory checks repeatable enough to survive disagreement.

A classification table both sides can apply

Severity classAppearance examplesHow the verdict uses it
CriticalSharp flash on handles, flaking coating on food-contact parts, missing safety componentsAny single case normally rejects the lot pending correction
MajorConspicuous scratches on display faces, misaligned prints, strong persistent odorCounted against the major limit; excess fails the lot
MinorFaint marks on hidden faces, slight thread unevenness, tiny bubbles inside molded partsCounted against the minor limit; excess fails the lot
ObservationSlight sheen variation within shade band, harmless mold-line tracesRecorded without counting; signals drift to watch
Packaging remarkScuffed retail box corners, faded carton print, loose inner fitCorrected before release; tracked separately from product defects

Circulate this table with the borderline photographs before production starts. When the factory team can classify findings the same way the inspector will, preparation replaces argument and first-pass rates rise across the season.

Illustration: A classification table both sides Decorative illustration for the section "A classification table both sides"; visual only, carries no data.

Turn findings into a lot verdict with sampling

Individual defects do not decide shipments; sampled defect counts against agreed limits do. The inspector examines units drawn at random from across every finished carton and records how many fall into each severity class, a procedure whose statistical foundation is described in acceptance sampling literature on judging lots from samples[1]. The report then compares those counts with the accept and reject numbers written into the purchase order, producing a verdict per class that any reader can retrace.

Sellers sometimes ask why a lot with visible flaws still passes, or why a mostly clean lot fails on a handful of findings. Both outcomes follow from the same arithmetic: thresholds judge rates, not anecdotes. A few minor marks spread thinly across a large sample stay within limits, while a cluster of major defects in the same sample exceeds them. Understanding this logic before inspection day prevents the emotional re-litigation of verdicts that the numbers already settled.

Many buyers coordinate this stage with UTS Inspection UTS Inspection, asking the team to verify workmanship findings against the specification independently before any rework decision. Independent verification matters most when the factory disputes the classification, because a second coded record with photographs settles whether the finding is major or minor. The cost of that clarity is small beside the cost of reworking the wrong defect class across thousands of units.

Keep critical findings outside the arithmetic

Critical defects sit outside the counting logic entirely. A single safety hazard, such as an accessible sharp point on a toy or exposed wiring on an appliance, normally stops the release regardless of how clean the rest of the sample looks. State this rule explicitly in the specification so that nobody treats a critical finding as one more entry in the major column.

Watch the observation column across seasons

Observations that never count still deserve trend review. A sheen variation recorded as an observation on three consecutive shipments signals process drift that will eventually cross into minor or major territory. Sellers who chart observations across purchase orders catch that drift while correction is still a conversation rather than a sorting operation.

Let destination rules inform severity

The same physical flaw can carry different weight in different markets, because labeling, safety, and consumer-protection rules differ by destination. A missing warning statement that is a packaging remark in one market may be a major finding where that warning is required, and an age-grade misprint moves from cosmetic to critical where toy regulations apply. Severity should therefore be set per destination rather than copied blindly between shipments.

Illustration: Let destination rules inform Decorative illustration for the section "Let destination rules inform"; visual only, carries no data.

This variation is normal rather than arbitrary. The international framework on technical barriers to trade describes how members pursue legitimate objectives such as safety and consumer information through regulations, standards, and conformity procedures[2]. For the seller, the lesson is administrative: record which severity assignments come from destination rules and which come from brand preference, so future shipments to new markets can be re-graded without rewriting the entire criteria set.

In practice, maintain one master criteria document with destination-specific severity columns. When a product line expands from one marketplace region to another, the team reviews only the affected rows instead of renegotiating every defect description. That structure also helps the inspector, who can see at a glance why a warning-text finding counts as major for this particular shipment.

Before approving mass production, confirm three things in writing: the borderline photographs for each defect type, the severity class of every destination-required marking, and the accept and reject numbers per class. These three attachments decide more verdicts than any other page of the purchase order.

Sources

  1. Wikipedia: Acceptance sampling —— Overview of acceptance sampling including AQL concepts used in product inspection.
  2. WTO: Agreement on Technical Barriers to Trade —— WTO portal for the TBT Agreement covering technical regulations, standards and conformity assessment.

Frequently asked questions

How detailed should appearance criteria be?

Detailed enough that two strangers reach the same verdict: state the defect, the measurement or viewing condition, and the threshold, then attach borderline photographs. Criteria that depend on adjectives alone, such as obvious or excessive, collapse at the first disagreement. Measurable wording plus photos is the standard that survives inspection day.

Who decides whether a defect is major or minor?

The purchase order decides, before production begins. The inspector classifies findings against the agreed table and photographs, and the report shows the classification for each finding. If the factory disagrees, the dispute is about the criteria document both sides signed, not about the inspector's opinion, which keeps resolution factual and fast.

Why can a lot with visible flaws still pass inspection?

Because acceptance thresholds judge defect rates within a random sample, not the existence of any single flaw. A few minor marks spread across a large sample stay within limits, while clustered major defects exceed them. The verdict reflects the agreed arithmetic applied to sampled evidence, which is exactly what makes it defensible to both sides.

Do severity classes change between destination markets?

They can, because labeling, safety, and information rules differ by destination. A missing warning may be a remark in one market and a major finding where that warning is required. Maintain severity per destination in the master criteria document, and mark which assignments come from market rules so expansion to new regions only requires reviewing the affected rows.