Quality Control and Inspection

Inspection built into the programme rather than applied at the end: staged checks, agreed defect classes, and a sealed sample everything is judged against.

Quality Control and Inspection is the set of checks applied to a leather bag programme from material intake through to container loading. Finished lots are inspected to AQL 2.5 under ISO 2859-1, single sampling, general level II, with defects classified critical, major and minor. The purpose is not to catch faults at the end but to agree, in advance, the arithmetic by which a lot passes or fails, so neither side is arguing from impressions. Physical testing against ASTM methods and third-party inspection are available on every programme.

What the Service Covers

Inspection runs at four points rather than one. A fault found at intake costs a hide; the same fault found at final audit costs a container. The stages below are ordered by how cheaply a problem can be caught.

How It Works

The inspection plan is written with the specification, not after it. Defect classes and tolerances are agreed before cutting begins, which is what makes a rejection defensible.

1. Specification review — the tech pack is reviewed and every measurable point is listed as an inspection criterion with a tolerance.
2. Defect classification — critical, major and minor classes are assigned so the acceptance arithmetic matches the brand's risk tolerance.
3. Material intake inspection — leather, hardware, lining and trims are inspected on arrival and rejected before they reach the cutting table.
4. Inline checks — inspectors sample at each critical process point and corrective action is applied immediately rather than at audit.
5. Final random inspection — the lot is sampled to AQL 2.5 under ISO 2859-1, single sampling, general level II, and the result recorded.
6. Pre-shipment audit and release — packing, marks and loading are verified, the report with photographs is issued, and shipment is authorised.

What You Provide / What We Provide

Inspection is only as good as the specification behind it. Where a brand has no in-house quality documentation, the programme team writes the inspection plan as part of development.

Defect Classes and Acceptance

Defects are graded before inspection, not during it. Critical defects are those that make a unit unsafe or illegal; major defects are those a customer would return the bag over; minor defects are those a customer would notice but not act on.

Timelines and Terms

Inspection is scheduled into the programme rather than added to it. Final inspection normally occupies three to five working days between production completion and shipment release.

Standards and Compliance

Inspection answers whether the goods match the specification. Compliance answers whether the specification is lawful in the destination market. Both are needed, and they are documented separately because a compliant specification can still be badly made.

• Inspection standard — AQL 2.5 under ISO 2859-1, single sampling, general level II, which fixes both the sample size drawn from a lot and the defect count at which the lot passes or fails.
• Management systems — work runs to ISO 9001 discipline with environmental aspects under ISO 14001 practice, on an SGS-verified 4,950 m² production floor in Quanzhou, Fujian; 137 staff, 7 lines, 149 machines.
• Chemical compliance — REACH declarations for the European market, CPSIA and lead testing for the United States, Proposition 65 for California, and OEKO-TEX certified linings on request.
• Physical testing — abrasion, seam strength, handle and strap load, zip cycling and colour fastness to rubbing, light and water, reported against ASTM methods.
• Hardware testing — saline spray, pull strength, zip cycle testing and nickel release for European compliance.
• Traceability — sealed samples, approved lab dips, inspection reports and test certificates retained against the style for re-orders and audit.

Writing the Inspection Plan Before Cutting Begins

An inspection plan is a list of things that can be measured, each with a method and a tolerance. Its value is that it turns a disagreement about quality into a measurement. Where a brand has its own quality documentation, the plan is written against it; where it does not, the plan is drafted from the tech pack during development. Either way it exists before cutting starts, because a criterion added after production is not a criterion at all — it is an opinion about finished goods.

The difficult part is usually not the criteria but the tolerances. A tolerance of two millimetres on a panel seam means something; an instruction that a seam should be neat does not, and two inspectors holding the same bag will reach different conclusions. The practical rule is that anything measurable is given a number, and anything that cannot be measured is judged against the sealed sample. That leaves a small residue of genuinely subjective points, which is precisely what a sealed sample exists to settle.

• Measurable points carry numbers — dimensions, stitch counts and cycle counts are written with a tolerance.
• Subjective points carry a sample — appearance is judged against a physical reference rather than a description.
• Zero-tolerance points named — the brand states which failures are unacceptable whatever the count.
• Method recorded with the criterion — how a point is checked forms part of the plan, not the inspector's discretion.
• Plan signed before cutting — agreed in writing so the assessment does not vary by shift or by inspector.

Brand-specific points deserve a line of their own. A mark that must sit a defined distance above a base seam, a lining that must not show at the mouth of a pocket, a puller that must fall to a particular side: these are the details a customer notices and a general checklist will not contain. Writing them down is what converts a preference into a criterion, and it is the only way such details are held consistently across a run and then across a season.

Where Deviations Originate on the Line

Every inspection report describes an outcome; the useful question is where the deviation entered. Most defects in leather bag assembly originate at four or five process points, and each leaves a characteristic signature. Recognising the signature is what turns a failed lot into a correction rather than a re-sort, because re-sorting removes the symptom and leaves the cause in place for the next run.

What distinguishes a well-run line is not the absence of these deviations but where they are found. A cutting error caught at the cutting table costs a panel; the same error found at final audit costs every unit assembled around it. That is the whole argument for inline checks, and it is why the inspection plan places a check at each critical operation rather than concentrating effort at the end of the run.

When a lot does fail final inspection, the report is written to identify origin rather than merely to count defects. The defect class tells the brand how severe the failure is commercially; the process point tells the line what to change. Where the same point appears across consecutive runs, the response is a specification or tooling review rather than another re-sort, because a recurring signature is a design problem rather than an operator problem.

Tolerances, Measurement and the Sealed Sample

Two kinds of judgement run through every inspection, and they behave differently. The first is measurement: a dimension falls inside the stated tolerance or it does not, and there is nothing to argue about. The second is appearance: whether an edge finish is acceptable, whether two panels read as the same colour, whether a mark sits where it should. Measurement is settled by the plan. Appearance is settled by the sealed sample, which is why sealing it properly is the most important single step in development.

The sealed sample also has to be the right thing to seal. It must be built in intended bulk materials and finish rather than in laboratory substitutes, because a sample built from a better hide than the run will use is a standard the run can never meet. It must be retained rather than shipped, and it must be the same physical object at final inspection as it was at sealing — a sample that has been handled, stuffed and carried for a season is no longer a reference for anything.

• Built in bulk materials — a sample in substitute materials sets a standard production cannot meet.
• Retained rather than shipped — the sealed sample stays against the style through the season and its re-orders.
• One physical object — photographs carry neither hand feel, edge finish nor true shade.
• Bench lighting agreed — appearance is judged under a defined light rather than whatever the room provides.
• Packing reference sealed as well — one packed carton is kept so presentation has a standard too.

Where a buyer and a programme team disagree at audit, the cause is nearly always a missing reference rather than bad faith. Either the tolerance was never written, or the sample was never properly sealed, or the packing reference was never kept. All three are inexpensive to fix before production begins and impossible to fix once it has finished.

Using Inspection Records Across Runs

Inspection produces a record, and the record is worth more than the release decision it supports. A single report tells a brand whether one lot passed. A sequence of reports tells it whether a style is stable, which process points drift, and whether a specification change made things better or worse. Any programme expecting to re-order should read that sequence, because the second run is the only place the first run's data can be used.

1. Report retained against the style — defect counts by class, with photographs, filed by lot and by run.
2. Defects grouped by process point — where the deviations originated matters more than how many there were.
3. Recurring signatures reviewed — the same point appearing twice is a specification or tooling question.
4. Change recorded with its outcome — a revision is filed alongside the report from the run that followed it.
5. Re-order planned against the trend — the next run is scheduled with the previous run's data in hand.

The practical payoff falls on the re-order. Because the pattern, tooling and approved materials are retained, a repeat run carries no sampling round, which means the only thing standing between it and a clean result is the correction identified the previous time. Where a corrective action was recorded and never applied, the same defect class reappears at the same rate. Where it was applied and verified on the line, the re-order usually improves without any change to the specification.

Third-party inspection belongs inside the same record rather than outside it. SGS, Bureau Veritas, Intertek or a buyer-nominated inspector are given the sealed sample, the approved dips, the previous report and the inspection plan, so the assessment is made against identical references. Where a third-party result differs materially from an internal one, the difference is worth investigating rather than averaging: it usually means the two inspections worked from different references.

Agree an inspection protocol

Send the specification and the market you are shipping to. The programme team will return an inspection plan, the defect classification and the documentation set that ships with your goods.

Request a QuoteWhatsApp +86 177 5002 0688Email service@junyuanbags.com

Frequently Asked Questions

What does AQL 2.5 actually mean in practice?

It is the acceptance threshold under ISO 2859-1. The standard fixes how many units are drawn from a lot at general level II and how many defects of each class the lot may contain before it fails. At AQL 2.5 with single sampling, a 500-piece lot draws 50 units, accepts up to three major defects and rejects at four. Both sides work from the same arithmetic rather than from impressions.

When does inspection take place during production?

At four points: incoming material, in-process at critical operations, final random inspection once production is complete, and a pre-shipment audit covering packing, marks and loading. Final inspection normally occupies three to five working days between production completion and shipment release. Inline checks are where most deviations are caught, because that is where correction is cheapest.

How are defects classified?

Critical defects make a unit unsafe, illegal or liable to regulatory action, such as nickel release above the permitted limit or a failed load-bearing strap. Major defects are those a customer would return the bag over, such as a broken stitch run, misaligned panel or failed zip. Minor defects are noticeable on close inspection but not grounds for return, such as slight shade variance between panels.

Can we nominate our own inspection company?

Yes. SGS, Bureau Veritas, Intertek or any buyer-nominated inspector can attend, and full documentation support is provided: the sealed reference sample, the approved lab dips, the material declarations and the previous lot's inspection report. Third-party inspection is welcome at any stage rather than only at final audit.

Which physical tests can be arranged?

Abrasion resistance, seam strength, handle and strap load, zip cycling, and colour fastness to rubbing, light and water, all reported against ASTM methods. On hardware, the usual set is saline spray, pull strength, zip cycle testing and nickel release for the European market. Results are issued before shipment when testing is commissioned.

What compliance documentation ships with each order?

Each shipment carries its inspection report with photographs, material declarations covering REACH for the European market or CPSIA for the United States as the destination requires, test certificates where testing was commissioned, the commercial documentation set and the certificate of origin. Proposition 65 and OEKO-TEX references are added where specified.

What happens if a lot fails final inspection?

The lot is not released. The report identifies which defect class caused the failure and where in the process the deviation originated, and corrective action is applied before re-inspection. Because inline checks run throughout production, a failure at final audit usually indicates a process drift rather than a specification problem.

How is the reference sample used during inspection?

The sealed pre-production sample is the standard. It is built in intended bulk materials and finish, retained against the style, and used to judge both the bulk run and the final inspection. Inspectors compare units against it rather than against a written description, which is why sealing it properly matters.

Is inline inspection independent of the production line?

Inspectors carry out random checks at critical process points and any deviation from the specification triggers corrective action on the line. The inspection plan, defect classification and acceptance criteria are all agreed in writing before cutting begins, so the assessment does not depend on who is on shift.

Do inspection standards change for small batch orders?

No. Volume does not alter the standard. A pilot run is inspected to AQL 2.5 under ISO 2859-1, single sampling, general level II, and smaller lots actually draw a proportionally higher share of units under the level II table. Coverage is not relaxed to accommodate a reduced quantity.

Who sets the tolerances used in the inspection plan?

They are agreed jointly before cutting begins. The brand states what it can accept at retail; the programme team states what the construction can hold repeatedly. Where those two do not meet, the specification is adjusted rather than the tolerance quietly widened at audit. Anything written into the plan is then checked by the recorded method.

Can we supply our own inspection checklist or quality manual?

Yes, and it is often the better starting point. A brand's own manual is written into the inspection plan, with each of its criteria given a method and a tolerance where one is missing. Where the brand has no quality documentation, the plan is drafted from the tech pack during development and issued for sign-off.

How is shade variation between panels judged?

Against the sealed lab dip and the sealed pre-production sample, under agreed bench lighting rather than the light the room happens to provide. Because no dye lot repeats exactly, an acceptable range is recorded at approval. Panels outside that range are classified by defect class rather than debated individually.

What is a zero-tolerance point?

A failure the brand will not accept at any count, regardless of what the AQL arithmetic would allow. Typical examples are a failed load-bearing strap, a mark position that breaks a brand rule, or a compliance failure such as nickel release above the limit. These are named in the plan before cutting and checked separately.

Are inspection records kept after the shipment has gone?

Yes. Reports with photographs are retained against the style, grouped by defect class and by process point, together with the sealed sample and approved dips. On a re-order those records tell the line which correction still needs verifying, which is usually the difference between a repeat run that improves and one that repeats the same result.