Before authorising bulk on a private label bag, inspect four things: the approved physical sample against the written specification, the actual materials and hardware against what was specified, the construction details that predict field failure, and the finished unit's function under load. Then record the decision against a documented standard — inspection to AQL 2.5 under ISO 2859-1, single sampling at general level II — and keep one dated reference unit. Release without those records is an assumption, not a control.
This article sets out the three gates a programme should pass before bulk begins, the physical checks that matter on the sample, how to verify materials and hardware against the specification, which construction details predict whether a bag survives use, what function and wear testing is proportionate, how defects are classified and counted under AQL 2.5, and how to document release so the decision is defensible months later.
The Three Gates Before Bulk
Quality in a bag programme is decided at three points, and only one of them looks like inspection. Understanding what each gate is for prevents the common error of inspecting hard at the end and discovering a problem that originated in a specification nobody wrote down.
| Gate | What is being checked | Who decides | What failure costs |
|---|---|---|---|
| Specification gate | The tech pack is complete, unambiguous and measurable | Brand owner | Everything downstream; defects become matters of opinion |
| Sample gate | The physical unit matches the specification | Brand owner, on the sample | One round of sampling, 6–10 working days |
| Pre-production gate | Bulk materials, hardware and setup match the approved sample | Programme team with the brand | An entire run made from the wrong input |
| Release gate | Finished goods meet the agreed standard | Inspector against AQL 2.5 | Shipment-level rework or rejection |
The specification gate is the one most often skipped and the one that determines everything else. A defect can only be called a defect if a measurable standard existed first. Without tolerances, without named materials and without a reference for colour, an inspection becomes a negotiation, and negotiations at the release gate are settled under time pressure.
Tolerances deserve a note, because new brand owners either specify none or specify them unrealistically. A bag is assembled from a natural material by hand-guided machines; asking for zero variance invites either an inflated price or a supplier who agrees and then ignores the clause. State the tolerance that reflects what the construction can hold and what the customer would notice, and hold to it.
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 to Check on the Physical Sample
The sample is the only fully controllable unit in the programme, and it should be examined slowly and in good light. Work through the bag in a fixed order so nothing is skipped, and record measurements rather than impressions.
| Area | What to examine | Common fault | How to record it |
|---|---|---|---|
| External dimensions | Every stated measurement, with the bag empty and lightly filled | Drift of several millimetres from the spec | Measured table against the tech pack |
| Panel alignment | Seams meeting at corners, symmetry left to right | Asymmetric gussets or skewed panels | Photograph both sides together |
| Stitch line | Density, straightness, distance from the edge | Wandering line, uneven spacing | Stitches per inch counted at three points |
| Edge finish | Coverage, smoothness, adhesion at corners | Thin or cracked edge paint | Close photograph of each corner |
| Hardware | Finish, function, attachment security | Plating marks, stiff zip, loose rivet | Cycle each fastener and note resistance |
| Lining | Fit, cleanliness, seam allowance inside | Pooling, twisting, visible adhesive | Inspect turned inside out |
| Branding | Placement, depth, legibility | Off-centre or shallow impression | Measured distance from a fixed datum |
Measurements taken with the bag lightly filled are more informative than those taken empty. A bag designed to hold a laptop, a wash bag or a day's worth of items behaves differently under load, and faults in gusset geometry and strap attachment only appear when the structure is stressed.
Verifying Materials and Hardware Against the Specification
A sample can be made from exactly what was specified while a bulk run is not. Material substitution is the single most common cause of a first bulk shipment disappointing a brand that approved a good sample, and it is prevented by naming things precisely rather than descriptively.
Where compliance matters, verify documentation at the same time. Material declarations supporting REACH conformance in the European Union, OEKO-TEX certification for textile components, and Leather Working Group evidence for tannery origin should be collected during sampling, not requested after a customs or marketplace query arrives.
Ask for a material sample card to be retained alongside the golden sample. A small swatch of the approved hide, lining and hardware finish, dated and filed, resolves most substitution disputes in a single email, because it converts a recollection into an exhibit.
Construction Details That Predict Field Failure
Some faults are cosmetic and some are structural, and the distinction matters because only one type generates returns. The table below lists the construction points that predict whether a bag survives its first year, in rough order of how often they cause warranty claims.
| Detail | Why it fails | Acceptable construction | Severity if wrong |
|---|---|---|---|
| Strap and handle attachment | Load concentrates at the join | Reinforced with a return stitch or rivet through a reinforcement patch | Critical: the bag becomes unusable |
| Gusset join | Stress at the base corner | Continuous stitch line with adequate allowance | Major: opening or tearing under load |
| Zip insertion | Repeated cycling pulls the tape | Tape captured on both sides, bar-tacked at each end | Major: zip separates from the panel |
| Stitch density | Too few stitches per inch tears progressively | Consistent density appropriate to the material weight | Major: seam failure in use |
| Edge paint adhesion | Flexing cracks a brittle finish | Flexible coating applied in thin, layered passes | Minor to major depending on area |
| Lining attachment | Poor fit causes pooling and tearing | Lining cut to the bag, not bagged in loosely | Minor: cosmetic but visible |
| Base reinforcement | Contents load the base continuously | Stiffener or double layer where the spec calls for it | Major: sagging and loss of shape |
Handle attachment deserves particular attention relative to its cost. It is the highest-loaded join on almost every bag, it is visible, and reinforcing it properly adds very little to the unit price. Buyers who compare quotes across suppliers frequently find that the saving was taken here, in a place no photograph reveals.
Function and Wear Testing That Is Proportionate
Full laboratory testing is unnecessary for most programmes, but a short list of practical tests catches most field failures before they reach a customer. None requires equipment beyond a scale, a load and patience.
- Cycle every fastener. Run each zip, clasp and buckle through fifty cycles and note any change in resistance; stiffness usually precedes failure rather than following it.
- Load the handles. Fill the bag to its intended working weight, lift and carry it, then inspect every attachment point for movement or distortion.
- Wear the strap. A crossbody strap should be worn loaded for an extended period; pressure points and edge comfort only appear in use.
- Tug the hardware. Pull each rivet, screw and D-ring firmly and check for rotation or movement in its seating.
- Check zip alignment. A zip that closes under tension will not survive; test it both empty and filled.
- Rub the edges. Flex a painted edge repeatedly and look for cracking or lifting at the corners where stress concentrates.
- Test the lining under load. Load the bag and examine whether the lining distorts, pools or pulls at its seams.
- Assess odour. A strong solvent smell in a sealed carton indicates insufficient cure time and frequently generates complaints on arrival.
Where a claim will be made in marketing, testing should match it. A bag described as water resistant should be tested as such, and a strap described as load-rated should be tested to that load. Trade estimates of durability mean little; a simple documented test of your own is worth considerably more in a dispute.
Defect Classification and the AQL Decision
Inspection to AQL 2.5 under ISO 2859-1, single sampling at general level II, is the working standard for bag programmes. The number sets how many units are drawn from a lot and how many defects are tolerated before the lot fails. It does not, on its own, say what counts as a defect — that comes from classification.
| Class | Definition | Examples on a bag | Typical tolerance |
|---|---|---|---|
| Critical | Makes the product unsafe or unusable, or breaches regulation | Broken strap attachment, sharp hardware edge, missing required labelling | Zero acceptance |
| Major | Affects function, saleability or service life | Failed zip, cracked edge paint across a seam, panel misalignment | Governed by the AQL level |
| Minor | A deviation noticeable on close inspection but not affecting use | Light surface mark, slight shade variation within tolerance, loose thread | Higher tolerance than major |
Write the classification into the specification before inspection, with examples specific to your product. Two inspectors reading the same standard will classify differently; two inspectors reading a standard with photographs of each class will largely agree. The photographs are the valuable part.
It is also worth agreeing what happens on a failed lot. Options include full re-inspection after rework, a second sampling under tightened inspection, or acceptance with a price adjustment for a defined minor-fault rate. Deciding this during the dispute, with freight booked, is the worst time to negotiate it.
Documenting Release: Golden Sample and Records
The output of a control process is a record. Without one, a programme cannot answer the question that arises in every second season: is this reorder the same as the original?
- Seal and date one approved unit as the golden sample, and store it away from light and handling.
- Retain a material card with swatches of every approved material, hardware finish and thread colour.
- File the final tech pack with revision history, so any later change is traceable to a decision.
- Keep the pre-production sample photographs, taken under consistent lighting and from fixed angles.
- Store the inspection report against the purchase order, including the sample size drawn and the counts by class.
- Record any accepted deviation in writing, with the reason and the price adjustment if one was agreed.
- Note the production slot and material lot references, which makes a later substitution observable.
Two copies of the golden sample are better than one. Hold one at the origin and one with the brand or its inspection agent, so that a dispute can be resolved by comparing two physical units rather than by exchanging photographs of one.
Remote and Third-Party Inspection
Most brand owners do not attend inspection in person, and the arrangement works well when the inputs are right. The inspector needs a specification, a golden sample and a classification guide; given those, a remote report is as reliable as an attended one and considerably cheaper.
| Approach | Cost | Reliability | Best used when |
|---|---|---|---|
| Attended inspection by the brand | High | Highest, but subjective without a spec | First run of a complex style |
| Independent third-party inspection | Moderate | High with a good specification | Most bulk releases |
| In-house inspection reported remotely | Included | High, if the report is detailed and photographic | Routine reorders of proven styles |
| Live video walkthrough | Low | Moderate; good for questions, weak for detail | Interim checks during production |
| Photographic approval only | Lowest | Low; misses feel, smell and function | Never sufficient on its own |
Ask for the report to include photographs of the faults found, not only the counts. A report stating four major defects tells you whether the lot passed; a report showing four photographs tells you whether the fault will recur, which is the more useful question for a reorder.
Timing the inspection correctly is the other half of the arrangement. Inspection should happen when the lot is sufficiently complete to be representative — typically when goods are packed or close to packed — and early enough that a failure still leaves room for rework before the booked sailing. An inspection conducted on a partial lot produces a result that cannot be applied to the whole.
Finally, treat inspection as feedback rather than as an audit. A recurring minor fault is information about the specification or the construction, and the appropriate response is to change the specification rather than to reject lots repeatedly. Programmes that improve their tech pack after each run need progressively less intervention at the release gate.
Set your release standard
Send your tech pack, target market and launch date. The programme team in Quanzhou will return an inspection checklist, a defect classification for your style and the AQL sampling plan for your quantity.
Frequently Asked Questions
What should I inspect before approving bulk bag production?
Four things: the approved physical sample against the written specification, the actual materials and hardware against what was named, the construction details that predict field failure, and the unit's function under load. Then record the release decision against a documented standard and keep a dated reference unit.
What is a golden sample and why does it matter?
A golden sample is one approved unit, sealed and dated, retained as the physical reference for the style. It settles whether a reorder matches the original, which is the question that arises in almost every second season. Holding a second copy at origin allows disputes to be resolved by comparing units rather than photographs.
How do I stop materials being substituted in bulk?
Name materials precisely rather than descriptively: tannery or mill, article reference, finish, thickness and colour reference for leather; catalogue number, base metal and plating thickness for hardware. Retain a dated material card with swatches, and record the lot references used for bulk so a substitution is observable.
What does AQL 2.5 mean in practice?
AQL 2.5 sets the acceptable quality limit used with ISO 2859-1 single sampling at general level II. It determines how many units are drawn from a lot and how many defects are tolerated before the lot fails. It does not define what a defect is, so pair it with a written classification.
How are bag defects classified?
Critical defects make a product unsafe, unusable or non-compliant and carry zero acceptance. Major defects affect function, saleability or service life. Minor defects are noticeable on close inspection but do not affect use. Write the classification with photographs of each class before inspection begins.
What is the most common cause of a disappointing first bulk run?
Material substitution after a good sample was approved. Descriptive specifications such as soft black leather allow a wide range of inputs, all of which satisfy the description. Naming the tannery, article reference, finish and thickness, and retaining a swatch card, removes almost all of the ambiguity.
Should I attend inspection in person?
For a first run of a complex style, yes if the budget allows. For routine reorders, an independent third-party inspection against a good specification, a golden sample and a photographic classification guide is as reliable and far cheaper. Photographic approval without physical inspection is never sufficient on its own.
What function tests should I run on a bag sample?
Cycle every zip, clasp and buckle fifty times; load the handles to working weight and inspect attachment points; wear a loaded crossbody strap; tug each rivet and D-ring; test the zip both empty and filled; flex painted edges; load and inspect the lining; and assess odour in a sealed carton.
How do I check stitch quality?
Count stitches per inch at three points and compare them with the specification, then check the distance from the stitch line to the edge and whether the line runs straight. Consistency matters more than a high count: a dense but wandering line indicates a handling problem that will vary across the run.
What is a pre-production sample?
It is a unit made from the actual bulk materials, with bulk hardware, on the production line rather than in the sampling room. It is the last point at which a material or setup error can be caught cheaply, and it should be approved before cutting begins rather than after.
How long does sampling take before bulk can start?
Six to ten working days per round, or twelve to fifteen days for complex constructions. Most programmes need two rounds, so budget four weeks from a stable brief to an approved sample. Bulk then runs 35–50 days after sample approval, which is the figure to plan the launch around.
What should an inspection report include?
The sampling plan and sample size drawn, counts by defect class, the acceptance decision, and photographs of each fault found. Photographs are the valuable part: counts tell you whether a lot passed, while images tell you whether the fault will recur on the reorder.
Who decides whether a lot passes inspection?
The inspector applies the agreed standard, but the brand owner owns the acceptance decision. Where a lot fails, options include re-inspection after rework, a second sampling under tightened inspection, or acceptance with a price adjustment for a defined minor-fault rate. Agree which applies before inspection.
How do I verify hardware quality?
Specify base metal, plating type and plating thickness, then check the finish under consistent light for marks and colour consistency across the run. Cycle each fastener and test attachment security. Plating thickness is the specification most quietly reduced to find a saving, and it governs how the finish wears.
Is laboratory testing necessary for bags?
Full laboratory testing is rarely necessary, but targeted testing is worthwhile where a marketing claim is made or a regulated market is involved. Material declarations supporting REACH conformance, OEKO-TEX evidence for textile components and tannery traceability should be collected during sampling rather than after a query.
What causes edge paint to crack?
Usually a coating applied too thickly or without sufficient flexibility for the material, or inadequate cure time before packing. Corners concentrate stress, so inspect them first. Ask for thin layered passes with a flexible formulation, and check for lifting after repeated flexing rather than only on arrival.
How should I handle a failed inspection?
Establish the defect class and cause first, then decide between rework with re-inspection, a second sampling under tightened inspection, or acceptance with an agreed adjustment. Notify the freight forwarder immediately if the date is at risk, since a delay of days in the decision can cost weeks in the slot.
Does odour matter in a quality check?
Yes. A strong solvent smell in a sealed carton indicates insufficient cure time before packing and reliably generates customer complaints and returns. Assess it when the carton is first opened rather than after the units have aired, and specify a minimum cure period in the tech pack.
Can I approve a sample from photographs alone?
Not reliably. Photographs conceal handle feel, material hand, odour, function under load and dimensional accuracy. Video improves matters for specific questions but is weak on detail. Attend, send an inspector, or approve a physical sample sent to you, and reserve photography for recording what was approved.
How much does inspection add to a programme?
Third-party inspection is a modest fixed daily charge relative to the value of the goods, and it is the cheapest available protection against a shipment-level error. Against a bulk order of 500 pieces or more, the cost is typically a small fraction of the value at risk in a single failed lot.