Most quality disputes in the bag trade are not caused by bad workmanship. They are caused by vague expectations: a buyer approves a sample, receives bulk, and discovers that the two disagree on details nobody wrote down. An inspection checklist closes that gap by converting an approved sample into a set of measurable, station-by-station checks that can be applied identically to the first carton and the last. This article sets out how a B2B inspection checklist for luxury bags is structured, what each station should test, how defects are graded, and which report fields a buyer should insist on before releasing a shipment.
The checklist below assumes inspection against AQL 2.5 under ISO 2859-1, single sampling, general inspection level II, which is the contractual baseline built into the programmes we coordinate. It is written for brand teams, sourcing managers and third-party inspectors who need a document that survives contact with an actual shipment rather than a generic form.
A word on scope. Inspection verifies a finished batch against an agreed standard; it does not create quality that was never specified. The most effective checklists are written during sampling, while changes are still free, and then frozen at approval. Anything added after that point tends to be applied inconsistently and argued about later.
Programme parameters at a glance
| Parameter | Commitment |
|---|---|
| Minimum order quantity | 500 pieces per style / colourway |
| Sampling lead time | 6–10 working days (12–15 days for complex constructions) |
| Bulk production lead time | 35–50 days after sample approval |
| Inspection standard | AQL 2.5 (ISO 2859-1), single sampling, general level II |
| Incoterms | FOB Xiamen — EXW, CIF and DDP available on request |
| Payment terms | T/T 30% deposit / 70% before shipment; L/C at sight for container loads |
| Verified production floor space | 4,950 m² (SGS-verified) |
| Team | 137 staff across 7 production lines |
| Equipment | 149 machines including computerised saddle-stitch and edge-painting units |
| Monthly capacity | 200,000 pieces |
| Entity | Founded 2014; principal has worked in bag development since 2004 |
What a B2B inspection checklist is actually for
Three jobs. First, it establishes a shared vocabulary between the brand, the programme coordinator and whoever holds the gauge on the line, so that acceptable means the same thing to all three. Second, it converts subjective impressions into countable findings, which is what makes an accept or reject decision defensible. Third, it produces a record that supports a claim if a dispute arises after the goods have shipped.
- Shared vocabulary: every defect class is defined by an observable condition, not by an adjective.
- Countable findings: results are recorded as counts against sample size, so decisions are arithmetic.
- Defensible decisions: an accept or reject follows from the recorded counts and the agreed acceptance number.
- Traceable record: the report carries batch identity, sample size and findings, supporting later claims.
- Corrective loop: recurring minor findings point to a specification change rather than to stricter inspection.
That last point is the one most often missed. A checklist is a diagnostic instrument as much as a gate. If the same minor defect appears in three consecutive batches, tightening inspection will not remove it; changing the specification, the tooling or the material will.
Checklists also protect relationships. When the standard is written down and agreed in advance, a rejection is a factual outcome rather than an accusation, and the conversation moves immediately to corrective action. When it is not written down, every rejection becomes a negotiation about what was meant, and both parties leave the season with less trust than they started with.
Sampling frame: applying AQL 2.5 under ISO 2859-1
AQL stands for acceptance quality limit, the worst tolerable process average when a continuing series of lots is submitted for acceptance sampling. It is not a promise that a batch contains at most 2.5 percent defects. It is a statistical threshold expressed through a sampling plan: a sample size, an acceptance number and a rejection number derived from the lot size and the inspection level.
| Lot size (pieces) | Sample size | Accept | Reject |
|---|---|---|---|
| 501 to 1,200 | 80 | 5 | 6 |
| 1,201 to 3,200 | 125 | 7 | 8 |
| 3,201 to 10,000 | 200 | 10 | 11 |
| 10,001 to 35,000 | 315 | 14 | 15 |
| 35,001 to 150,000 | 500 | 21 | 22 |
Two practical notes. Sample units are drawn at random across the whole lot, including across cartons, not from the top of the first box, because defects cluster by production hour and by operator. And the plan must state which defect class it governs: most programmes apply a tighter limit to critical defects and a looser one to minor defects, rather than a single threshold to everything.
Sampling also has a psychological dimension worth planning for. An inspector who knows the acceptance number will count carefully; one who does not will sort by impression. Publishing the plan, including code letter, sample size and acceptance numbers, on the first page of the report keeps the exercise arithmetic rather than argumentative, and makes it possible for a second party to reproduce the result from the same lot.
Station one: incoming material and colour verification
Material is checked before a blade touches it, because every downstream defect is cheaper to prevent here than to sort later. The reference is the approved material swatch retained at sampling, held under the same light source used at approval.
- Confirm hide or roll identity against the approved swatch, including finish, grain character and hand feel.
- Compare colour under controlled lighting and record any deviation beyond the agreed tolerance.
- Check thickness with a gauge at several points on the sheet, not once at the edge.
- Verify backing adhesion on coated materials by flexing a corner and examining for lift.
- Confirm quantity and dye-lot identity, and record the lot reference on the inspection report.
- Reject material with scars, growth marks or loose grain break in zones that will appear on visible panels.
Colour deserves a specific tolerance rather than a judgement call. Without a number, every observer is right. With a number, and with both parties holding a swatch from the same dye lot, the conversation becomes short.
Material checks should also capture lot identity for traceability. If a complaint arrives months after delivery, the fastest route to an answer is knowing which hide or roll the affected units came from and whether other units in the same lot behaved differently. Recording that reference costs a line on a form and saves weeks of correspondence.
Station two: cutting, skiving and panel preparation
Cutting errors propagate invisibly. A panel cut two millimetres out of true still stitches together, but the finished bag sits asymmetrically, and the fault only becomes obvious at the assembly station when it is expensive to recover.
| Check | Method | Typical failure |
|---|---|---|
| Panel dimensions | Compare against the pattern template on a flat table | Progressive drift as blades wear |
| Grain direction | Visual check against the marked direction on the pattern | Panels cut across the grain, changing drape |
| Skiving depth | Gauge the thinned edge at several points along the seam allowance | Over-thinned edges that tear at the stitch line |
| Edge cleanliness | Run a finger along the cut edge under strong light | Ragged or compressed edges that finish poorly |
| Matching and symmetry | Lay left and right panels face to face | Visible asymmetry once the bag is assembled |
Station three: stitching and seam integrity
Stitch inspection covers geometry, tension and security. Geometry is measured, tension is assessed by feel and by the behaviour of the seam under load, and security is verified by examining the start and finish of every run.
- Count stitches per centimetre against the specification, at three positions on each seam type.
- Check seam allowance width for consistency along the full run.
- Examine thread tension: the stitch should sit on the material, not pull into it.
- Inspect start and finish: backstitching or a secured knot, with no loose tail on the outside.
- Pull gently along the seam at several points and look for thread grinning through the material.
- Verify that no stitch crosses or clips a previous one on overlapping runs.
Grinning thread, where the stitch sits proud and shows the thread beneath the surface when the seam is flexed, is one of the most common major findings in leather work and one of the easiest to correct at the machine before the bag is closed.
Thread itself should be recorded by reference rather than by colour name. Two spools described identically on a purchase order can differ in shade, twist and finish, and a seam sewn with a substituted thread may look correct on the day and fail months later under UV exposure or abrasion. The reference is what makes the substitution detectable.
Station four: hardware, closures and function
Hardware is where function and appearance meet, so it is checked twice: once as a component and once in operation. A zip that runs smoothly on the bench may bind once it is set into a curved panel, which is why functional testing belongs at this station rather than at final audit.
Record hardware findings by component reference rather than by description. A plated component can drift in shade between production runs, and only a recorded reference allows the buyer and the programme coordinator to establish whether the bulk components match the approved set.
Station five: assembly, shape, symmetry and interior
This is the station at which the bag becomes an object rather than a set of parts, and it is where most subjective judgements live. Convert them into measurements wherever possible: handle drop, base footprint, symmetry across the centre line, and the behaviour of the bag when set down empty.
- Set the bag on a flat surface and check that it stands without rocking or leaning.
- Measure handle drop on both sides and compare against the approved sample.
- Check symmetry across the centre line from front, back and base.
- Inspect the interior for lining slack, trapped adhesive and exposed raw edges.
- Confirm that pockets lie flat and that no interlining has shifted during closing.
- Verify brand marks, serial plates and care labels for placement, alignment and legibility.
Edge finishing deserves its own line on the report. Painted edges should be even in colour and thickness with no drag marks; turned edges should sit flat without bulk at the corners; bound edges should follow the contour without puckering. Because edge faults are visible in the hand, they disproportionately affect perceived quality relative to their cost to fix.
Station six: packing, labelling and carton presentation
Packing is a quality station, not an administrative one. Bags arrive creased, compressed or marked because of packing decisions made under time pressure, and the customer experiences that damage first.
| Check | Requirement | Why it fails |
|---|---|---|
| Protection | Each unit in a dust bag, with soft filler in voids | Compression creases during stacking |
| Carton strength | Board grade and stacking height agreed in the specification | Cartons crush in transit and mark the leather |
| Quantity accuracy | Count per carton verified, mixed sizes prohibited | Short shipments discovered at destination |
| Labelling | Style, colour, quantity and carton number legible | Warehouse receiving errors and delayed sell-in |
| Documentation | Packing list matches the physical carton count | Customs and reconciliation delays |
Where units ship directly to retail, add a presentation check: dust bag orientation, tissue fold, and the position of any card or care leaflet. Recipients read these as deliberate or careless within seconds, and a presentation fault is indistinguishable from a product fault in the customer's mind even though it costs a fraction to correct.
Defect classification: critical, major and minor
Grading is what turns a list of observations into a decision. Every finding is assigned to one class, the counts are compared with the acceptance numbers for that class, and the lot is accepted or rejected on the worst-performing class.
| Class | Definition | Representative findings |
|---|---|---|
| Critical | Renders the unit unsafe, unlawful or unsaleable | Broken handle attachment, sharp exposed metal, missing compliance labelling |
| Major | Reduces saleability or service life; the customer would return it | Visible grain damage on a front panel, closure failure, asymmetric shape, adhesive on a visible surface |
| Minor | A deviation from specification unlikely to prompt a return | Slight stitch pitch variation, a faint mark inside the lining, minor packing presentation issues |
Programmes typically apply the tightest acceptance limit to critical defects, AQL 2.5 to major defects, and a looser limit to minor defects. Whatever the numbers, they must appear in the contract alongside the sample size, because an AQL figure without a plan is not enforceable.
One grading rule prevents most disputes: assign each unit its worst defect rather than counting every defect on every unit. If a bag has three minor issues and one major issue, it contributes one major finding to the count. Counting everything inflates the numbers with units that would have passed, and produces rejections that are hard to defend.
The inspection report: fields to require
A report that cannot be acted on is decoration. Require these fields, and require them in writing before the inspection rather than after it.
- Purchase order, style reference and colourway, with the approved sample reference.
- Lot size, sample size, inspection level and the AQL applied to each defect class.
- Acceptance and rejection numbers actually used, taken from the relevant ISO 2859-1 table.
- Findings listed by station, each with defect class, count and photographic record.
- Measurements taken, with the specified value and the recorded value side by side.
- Carton count, packing condition and quantity verification result.
- A clear conclusion: accepted, accepted with conditions, or rejected, with the reason stated.
- Corrective action notes, including who is responsible and by when.
Photographic records deserve a rule of their own. Every finding should be photographed with a scale reference and a note of which unit it came from, because a photograph without context cannot be reviewed later and a finding without a photograph cannot be challenged.
Requiring a conclusion in one of three fixed states, accepted, accepted with conditions, or rejected, prevents the most common failure of inspection reporting: a long list of findings with no decision attached. Conditions should be written as actions with an owner and a date, and any lot released under conditions should be re-verified before it ships.
Fitting inspection into the programme calendar
Inspection sits between the end of bulk production and the release of the shipment, and its timing has to be booked rather than hoped for. Bulk runs take 35 to 50 days depending on construction complexity and order size; sampling takes 6 to 10 working days, or 12 to 15 for complex constructions. A pre-production sample made on the line, inspected before the bulk run begins, is the single most useful checkpoint in that calendar.
We coordinate inspection on our Quanzhou production floor, a 4,950 square metre SGS-verified site with 137 staff, seven lines, 149 machines and capacity for 200,000 pieces a month. Buyers are welcome to appoint their own third-party inspector against the same checklist; the value of a written specification is precisely that it makes the result the same whoever holds the gauge. Standard commercial terms are FOB Xiamen with T/T 30/70, and the quality control service page sets out what the inspection stage covers.
Turn your approved sample into an inspection checklist
Send your style, material specification and tolerance requirements, and the programme team will return a station-by-station checklist with defect classes and report fields ready to use. Email service@junyuanbags.com, or start a WhatsApp conversation on +86 177 5002 0688.
Frequently Asked Questions
What does AQL 2.5 actually mean for a bag shipment?
AQL 2.5 is an acceptance quality limit applied through a sampling plan, not a guarantee that no more than 2.5 percent of units are defective. Under ISO 2859-1 single sampling at general level II, the lot size determines a sample size, which then determines acceptance and rejection numbers. A lot passes when the count of defects in that class is at or below the acceptance number.
How many units are inspected in a 1,000-piece order?
At general inspection level II with an AQL of 2.5 for major defects, a lot of 501 to 1,200 pieces draws a sample of 80 units, with acceptance at five defects and rejection at six. Sample units are drawn at random across the whole lot and across cartons, because defects tend to cluster by production hour rather than distribute evenly.
Should critical, major and minor defects use the same AQL?
No. Most programmes apply the tightest limit to critical defects, AQL 2.5 to major defects, and a looser limit to minor ones. Using a single threshold for everything either lets serious faults through or rejects batches over cosmetic issues. Whatever is agreed must be written into the contract together with the sample size and the table used.
When should the checklist be written?
During sampling, while changes are still free. The approved sample is the reference for every later check, so the checklist should be built alongside it and frozen at approval. Adding criteria after bulk production starts almost guarantees inconsistent application and a dispute about whether the standard was ever agreed.
What is the difference between a pre-production sample and a bulk inspection?
A pre-production sample is made on production tooling before the run begins and reveals problems a hand-finished prototype hides. Bulk inspection verifies the finished batch against the agreed standard and produces the release decision. The first prevents defects; the second counts them. Relying on the second alone is far more expensive.
How is colour deviation measured objectively?
By comparing the production material against a retained swatch from the approved dye lot under a controlled light source, with a numerical tolerance agreed in advance. Without a number, every observer is right and no decision is possible. Recording the dye-lot reference on the report also lets later reorders be matched to the original colour.
What are the most common major findings in leather bag inspection?
Visible grain damage or scars on front panels, closures that bind or fail under cycling, asymmetric shape once the bag is assembled, adhesive visible on an exterior surface, and uneven or dragged edge paint. All five are detectable at intermediate stations, which is why station-level checks matter more than a final visual audit.
How should hardware be tested?
Twice: as a component and in operation. Check plating for pitting and consistent shade, confirm rivets and screws are seated and do not rotate, then cycle each closure ten times to catch binding caused by setting into a curved panel. Record findings by component reference so bulk parts can be compared with the approved set.
Why is packing treated as a quality station?
Because most transit damage originates there. Insufficient void filler produces compression creases, weak board crushes under stacking and marks the leather, and quantity errors surface only at destination. Packing checks cover dust bags, filler, carton grade, per-carton counts, labelling and whether the packing list matches the physical cartons.
Can a buyer appoint their own inspector?
Yes, and a written checklist makes that straightforward: the same specification, the same sampling plan and the same defect classes produce the same result whoever holds the gauge. Buyers should confirm in advance which ISO 2859-1 table, inspection level and AQL apply, and agree the report format before the inspection date.
What should an inspection report contain?
Purchase order and style reference, lot and sample size, the inspection level and AQL used for each defect class, acceptance and rejection numbers, findings listed by station with class and count, measurements against specification, carton and quantity verification, and a clear accept or reject conclusion with reasons.
How are photographic records best handled?
Every finding should be photographed with a scale reference and a note identifying the unit. A photograph without context cannot be reviewed later, and a finding without a photograph cannot be challenged or verified. Requiring both at the outset removes most arguments about whether a defect was real or how serious it was.
What happens when a lot is rejected?
The report should state the reason and the corrective action, including who is responsible and by when. Options are a full re-sort with re-inspection, rework of a specific defect, or, where rework is not viable, a partial shipment. Re-inspection uses the same plan and the same checklist, and its results are recorded separately.
Does inspection replace the need for a clear specification?
No. Inspection verifies a batch against an agreed standard; it does not create quality that was never specified. Unwritten expectations cannot be inspected for, which is why tolerances, materials, hardware references and finish standards all need to be recorded at sample approval rather than discovered at final audit.
How does the minimum order quantity affect inspection?
Programmes are structured around 500 pieces per style and colourway, which falls into the lowest sampling band at general level II. Because the sample size is a fixed count rather than a percentage, smaller lots are inspected at a relatively higher rate. Each colourway is treated as a separate lot for colour conformity purposes.
Who coordinates inspection within a bespoke bag programme?
A sourcing and programme-management company acting as programme coordinator prepares the checklist, books the inspection into the production calendar and holds the result against the approved sample. Inspection is scheduled after bulk production of 35 to 50 days and before shipment release. Contact service@junyuanbags.com or WhatsApp +86 177 5002 0688.