Table of Contents
- Executive Summary: What a Factory Visit Must Prove
- Documents to Request Two Weeks Before You Travel
- The Cutting Station: Fabric, Marker and Ply Control
- The Stitching Floor: What Each Machine Should Be Doing
- Inline QC and the AQL 2.5 Final Inspection Station
- 5S Housekeeping and Needle Control
- Machine Maintenance Logs and Spare Parts Discipline
- Red Flags During the Walk-Through
- Virtual Audits and the Written Action Plan
- Frequently Asked Questions

Executive Summary: What a Factory Visit Must Prove
A backpack factory visit is worth the trip when it answers three questions that no sample can: whether the line that produced your approved sample is the line that will run your bulk order, whether quality control is a staffed station or a clipboard, and whether the documents on the wall match the documents in the file. At a 500-piece minimum order quantity (MOQ) per style and colour, with bulk production running 35-50 days and sampling at 6-10 working days (12-15 for complex builds), one bad shipment costs far more than a return flight to Fujian.
This checklist is written for UK brand owners and buying managers who have never walked a bag line, and for the ones who have and want a repeatable scorecard. It covers the stations you must physically inspect, the inline and final inspection points that hold AQL 2.5 to ISO 2859-1, the housekeeping and maintenance evidence that predicts whether a factory can repeat a good sample, the documents to request before you fly, the red flags that justify walking away, and a virtual alternative for buyers who cannot travel.
Use it alongside a technical pack and an agreed inspection protocol, never instead of them. A visit confirms what a supplier says about itself; it does not replace a written specification, a signed gold sample and an independent inspection at dispatch.
Score every area from 1 to 5 and weight the result. A workable threshold is a weighted score of 3.8 or better out of 5, with no single area below its minimum pass mark.
| Area to score | What you are actually judging | Weight | Minimum pass |
|---|---|---|---|
| Cutting station | Fabric inspection, shade segregation, marker efficiency | 20% | 3 / 5 |
| Stitching lines | Machine mix, work aids, stitch density, line balance | 25% | 3 / 5 |
| Inline and final QC | Inspector ratio, defect definitions, AQL records | 25% | 4 / 5 |
| 5S housekeeping | Floor markings, WIP control, needle policy | 10% | 3 / 5 |
| Maintenance records | Schedule, log completeness, spare parts crib | 10% | 3 / 5 |
| Documentation | Licences, audit reports, test reports, traceability | 10% | 4 / 5 |
Documents to Request Two Weeks Before You Travel
The paperwork you request in advance tells you what to look for on the floor. Ask for it 10 to 14 days before the visit, in a single email, and note which items arrive late or arrive as a photograph of a screen rather than a scan.
- Business licence and export registration. Confirms the legal entity name matches the name on your proforma invoice and the name on the bank account you will pay.
- Third-party audit reports. Social compliance and technical audits from recognised firms, with the date and the corrective action plan attached. An audit report without a CAPA is a marketing document.
- Quality manual and procedures. You are looking for documented incoming inspection, in-process control, final inspection and non-conforming goods handling. A supplier whose quality management is aligned to ISO 9001:2015 procedures will produce a controlled document with revision numbers; see the BSI guidance on ISO 9001 quality management for what a documented system should contain.
- Machine list. Count of single-needle, double-needle, bartack, overlock, post-bed, cylinder-arm, hot-knife and ultrasonic machines, with ages. A 7-line facility able to run 500-piece orders in 35-50 days will have a machine mix that matches the construction in your tech pack.
- Maintenance schedule and a sample log page. If the schedule exists but the log is blank, the schedule is decoration.
- Calibration certificates. Tension gauges, weighing scales, illuminance readings and measuring tapes.
- Fabric and hardware test reports. REACH and substance testing for dyed fabric, coating and metal hardware, aligned with ECHA guidance on REACH obligations.
- Subcontractor list. Which operations, if any, leave the site: embroidery, printing, metal plating, zipper assembly.
Read the responses as evidence. A supplier that sends a scanned, dated, complete set within 48 hours is telling you how it will handle your production file.
Confirm the shipping basis in the same email. FOB Xiamen is the standard default for UK buyers, and a supplier quoting CIF or DDP should explain exactly what is bundled into the price.
Bring your own copies, plus a one-page summary of your programme: styles, quantities per colour, ship window and destination port. Compare what you receive against the wider supplier audit checklist before you decide what to verify in person.
The Cutting Station: Fabric, Marker and Ply Control
Cutting is where most hidden quality problems are created, because a mis-cut panel can be sewn perfectly and still fail at the end of the line. Spend 20 minutes here and ask four questions.
How is incoming fabric inspected? A serious operation inspects rolls against a four-point system on a light table or inspection machine, records roll width, shade and defects per 100 square metres, and quarantines rejects. Ask to see a completed inspection sheet for the fabric used on your style, not a blank template. With 420D to 900D polyester and nylon, you should also see checks for coating adhesion, backer condition and width tolerance, because a two-centimetre width shortfall changes marker yield on every ply.
How are shade lots segregated? Dye lots must be cut as a set and never mixed within one order without your written approval. Look for physical separation, coloured roll-end tags and a cut ticket that records the shade number against each bundle. If rolls from three lots are leaning against the same table with no tags, expect a two-tone shipment.
How is the marker laid? Ask for the marker efficiency figure on your style. Well-optimised backpack markers typically land between 85% and 90%; anything under 80% is either a badly nested pattern or a style with unusual panel shapes, and you are paying for that waste in the unit price. Ask how many plies are cut at once, and whether the fabric is relaxed for 24 hours before spreading to prevent shrinkage distortion after sewing.
How is the cut counted and traced? Every bundle should carry a cut ticket with style, colour, size, ply count, shade lot and quantity. Cutting machines - straight knife, band knife, die press or CNC cutter - should have a numbered log. Ask for the number of panels cut versus panels issued to stitching on yesterday's order; a gap of more than 1% needs an explanation.
The Stitching Floor: What Each Machine Should Be Doing
Walk the stitching lines without stopping at the first table. Count machines, count operators, and check that the construction on the floor matches the construction in your tech pack.
Machine mix. Backpacks need more than single-needle flat machines. Expect bartack machines for strap anchors and zip ends, double-needle or compound-feed machines for load-bearing seams, post-bed or cylinder-arm machines for gussets and curved pockets, overlock or binding stations for lining edges, and a hot-knife or ultrasonic station if your design uses cut-and-seal webbing or laminated panels. If a factory proposes a 900D PU-backed hiking pack but has no bartack machine visible, the reinforcement will be decorative.
Work aids and jigs. The single best predictor of consistency at 500 pieces is the presence of physical guides: folder attachments, edge guides taped to the table, pre-cut templates, marked seam allowance lines, and a first-off approved piece pinned above each station. Ask to see the first-off sample for your style and check it carries a signature and a date.
Stitch density and thread. Standard backpack seams run 7 to 9 stitches per inch; bartacks are denser and shorter. Ask what thread ticket is in use (typically bonded nylon or polyester, 40s or 20s for heavy seams) and whether needle size is specified per operation. A line running one needle size for everything is a line that has never been audited.
Line balance. Look at the piles. A balanced line has small, even buffers between stations. One operator sitting behind a tower of cut panels while the next station idles means the line is unbalanced, which produces both overtime and rushed seams at the bottleneck.
Bundle discipline. Every bundle should have its ticket, and no bundle should be on the floor. If you see loose panels without tickets, ask how the factory would trace a defect back to a shade lot or an operator.
On the floor, the test is simple: pick three random operators, ask each to explain the seam allowance and the next operation, and see whether the answers agree with the tech pack. Fabric choice drives what the line can do, and this comparison of 420D versus 600D versus 900D backpack fabric explains why.

Inline QC and the AQL 2.5 Final Inspection Station
Quality control is either a roving function on the line or a room at the end of it, and the difference shows up in your claim rate. In a well-run bag plant you should find inspectors walking the line at defined intervals, typically one roving inspector for every 30 to 40 sewing operators, checking at roughly 30%, 60% and 90% of the run rather than only at the end.
Ask to see the defect classification sheet. It should split defects into critical, major and minor, with photographs and written definitions: broken stitch, skipped stitch, needle hole, uneven topstitch, wrong thread colour, hardware burr, zip skip, reversed panel, incorrect label placement, excess thread tails. If the sheet is generic across all products, it will not catch the failures specific to your design.
The final inspection station should be a lit, dedicated area with a folding table, a measuring station and a written sampling plan. The standard most UK buyers should specify is AQL 2.5 per ISO 2859-1, single sampling, normal inspection, applied to major defects, with critical defects at zero tolerance. Under that plan a lot of 501 to 1,200 pieces is inspected at general inspection level II using sample size code letter J: 80 pieces pulled at random, accept on 5 or fewer major defects, reject on 6. The standard itself is published by ISO as 2859-1, and the sampling tables are what an inspector will actually work from.
On site, verify four things. First, that the inspector pulls samples from cartons at random rather than from a prepared pile. Second, that the inspection record captures carton numbers and quantities, so a failure can be traced. Third, that failed lots are physically segregated with a red tag, not quietly reworked into the same cartons. Fourth, that the same standard is applied to your pre-shipment inspection at dispatch, since a factory that passes its own goods at AQL 4.0 will not pass an independent AQL 2.5 check.
5S Housekeeping and Needle Control
Housekeeping is not about tidiness for its own sake. In a cut-and-sew plant it predicts contamination, mixed components, lost tools, fire risk and needle fragments in finished goods. Ask whether the facility runs a 5S programme - sort, set in order, shine, standardise, sustain - and then test the claim in five minutes.
- Floor markings. Aisles, WIP zones, reject zones and walkways should be marked and respected. If trolleys are parked across the evacuation route, the markings are paint, not a system.
- Shadow boards and tool control. Scissors, rulers, screwdrivers and gauges should have a labelled home with a visible gap when an item is missing.
- Waste handling. Fabric offcuts should be bagged or baled at the point of generation. Loose trimmings under tables migrate into lining and get shipped.
- Lighting. Inspection areas should be measurably lit; 500 lux at the inspection surface is a common target, and dim corners are where skipped stitches survive.
- Needle control. This is the one to test hardest. Ask for the broken needle log: date, machine number, operator, needle size, whether every fragment was recovered, and who verified it. Ask how many needles were issued this month versus how many broken needles were recorded. A plant that cannot reconcile the two has no needle policy, only a story.
Machine guarding, cable management and emergency stops are worth a glance: the attitude that ignores a frayed cable usually ignores a loose thread. Benchmark basic machine safety using HSE guidance on work equipment and machinery.
Machine Maintenance Logs and Spare Parts Discipline
A sewing machine that is out of time produces skipped stitches at a rate no inspector can catch on a 500-piece run. Maintenance records are therefore a leading indicator of quality, not an administrative nicety.
Ask for three artefacts. The preventive maintenance schedule, which should list daily actions (cleaning, oil level, needle condition), weekly actions (tension check, belt tension, lubrication points) and monthly or quarterly actions (timing check, motor service, safety device test). The completed log for the last 90 days, ideally signed per machine or per line. And the downtime record, which shows how quickly faults were cleared.
Read the log against the machines in front of you: if line 4 runs eight machines and shows no entries for machines 4-6 in three months, nobody is writing it down. Entries should carry a machine identifier matching a plate on the machine, not just a line number.
Then walk to the spare parts crib. A plant running 7 production lines cannot afford to wait two weeks for a bobbin case, so it holds consumables: needles in the specified sizes, bobbins, feed dogs, presser feet, tension springs, belts, blades for the cutting knife, and at least one spare servo motor. Ask what the longest recent wait for a part was. An answer in days is a managed risk; a shrug is not.
Needle and thread storage matters as much as machines. Thread should be stored away from direct sunlight and dust, with colour cards maintained so that a reorder at month three matches the lot at month one. Ask how the factory handles a mid-order thread shortage, because a substitute ticket is one of the most common causes of visible shade variation on black and navy bags.
Red Flags During the Walk-Through
Most bad factory visits are not dramatic. They are a series of small inconsistencies that a buyer explains away because the sample was good and the price was right. Treat the following as prompts to ask one more question, not as automatic disqualifiers.
- The sample room is impressive and the line is empty. Ask which live order is running today and see the production schedule on the wall.
- The QC records are all completed in the same pen and the same handwriting. Records filled in retrospectively are not records.
- One inspector covering 60 machines. Inline control has been designed out.
- No shade segregation. Rolls from different dye lots are mixed on the same table.
- Unexplained buildings or locked areas. Ask directly whether any operation is subcontracted and where it happens. Subcontracting is legal and normal; concealing it is not.
- Capacity claims that do not fit the floor. Count machines, count shifts, estimate output per machine, and compare with the number quoted to you.
- Rust, oil or dust on stored hardware. Zips, sliders, D-rings and buckles should be bagged and boxed, not open in a tray by a window.
- No calibration stickers. Measuring tapes and tension gauges without calibration cannot support an AQL claim.
- Reluctance to let you photograph. A reasonable NDA request is fine; a blanket refusal to photograph anything is a signal.
- Different entity names. The company that quotes, the company on the licence and the company on the bank account should be the same, or the relationship should be explained in writing.
The strongest single test is to ask for a short, unannounced walk from cutting to packing on a live order, with no preparation. Factories that are genuinely well run look the same on a Tuesday morning as they do on a hosted visit.
If you find two or more of the items above, convert them into written conditions and verify them at the pre-shipment inspection. A 500-piece run rarely justifies abandoning a supplier over housekeeping alone, but it does justify an independent final inspection.
Virtual Audits and the Written Action Plan
Not every UK buyer can fly to Fujian, and a well-structured virtual audit catches perhaps 70% of what an on-site visit catches. It works when the route is fixed in advance and the evidence is verifiable.
Ask for a live video walk-through on a fixed route - gate, fabric store, cutting, stitching, assembly, QC, packing, loading bay - with no cuts, pausing at each licence and log so dates are legible. Request timestamped 4K photographs at 12 to 15 checkpoints, a clip of the machine plates, and an operator of your choosing holding a ruler against a seam so you can count stitches per inch.
For higher assurance, commission a third-party pre-production audit and a pre-shipment inspection at AQL 2.5 per ISO 2859-1; the report is independent evidence you can rely on later. A virtual audit cannot smell solvent adhesive, feel coating hand or check ventilation, so accept that limit and price it in.
Whether you visit in person or on video, the output must be the same: a one-page report with scores by area, photo evidence, a list of non-conformities ranked by risk, and a corrective action plan with an owner and a date for each item. Send it within 48 hours, ask for written acceptance, and re-verify the open items at the pre-production meeting before bulk starts.
Then write the result into the commercial documents. The inspection standard, the sampling plan, the rework allowance and the approval rights belong in the purchase order, not in an email chain. Buyers who want the numbers on a first order can request a quotation structured line by line, and should read the pre-shipment inspection guide before agreeing terms. The company behind this guide, QUANZHOU JUNYUAN BAGS, was founded in 2014 with a founding team in the trade since 2004, and publishes its facility data - 4,950 m2, SGS-verified, 137 people, 7 production lines - precisely so that a visit or a virtual audit has something concrete to check against.

Frequently Asked Questions
How long should a backpack factory visit take?
Allow two to four hours for one product family, covering cutting, stitching, assembly, QC and packing. Book a full day for multi-line audits and ask to see a live order.
Can I visit a factory before placing a first order?
Yes, and most export-oriented bag plants welcome pre-order visits because they shorten the sampling cycle. Expect to pay a sampling fee of 80 to 150 US dollars per style, which is typically credited against the bulk order once you proceed.
What documents should I request before the visit?
Request the business licence, third-party audit reports with corrective action plans, the quality manual, the machine list, the maintenance schedule and a completed log page, calibration certificates, fabric and hardware test reports, and a subcontractor list. Send the request 10 to 14 days ahead and note which items arrive late.
Which station matters most during a walk-through?
The final inspection station, because it is where the factory's real standard becomes visible. If the sampling plan, defect definitions and segregation of failed lots are sound there, the rest of the plant usually follows.
What AQL should I specify for backpacks?
Specify AQL 2.5 per ISO 2859-1 for major defects, with critical defects at zero tolerance. For a lot of 501 to 1,200 pieces this means an 80-piece sample and acceptance on 5 or fewer major defects.
How can I check 5S housekeeping in five minutes?
Look for marked aisles that are actually clear, shadow boards with visible gaps for missing tools, offcuts bagged at the machine, and a broken needle log with every fragment accounted for. Any one of these missing suggests the programme is on a poster rather than on the floor.
What are the biggest red flags on a bag line?
QC records completed in identical handwriting, one inspector covering 60 machines, mixed dye lots on the cutting table, and capacity claims that do not match the machine count. Two or more of these justify tightening inspection rather than walking away immediately.
Can a virtual factory audit replace an in-person visit?
It can replace it for documentation, line layout, machine mix and housekeeping, which is roughly 70 per cent of the value. It cannot assess smell, ventilation, coating hand-feel or whether fire exits physically open, so pair it with a third-party inspection.
What is a needle control policy and why does it matter?
It is a written procedure for recording every needle issued and every needle broken, with confirmation that all fragments were recovered. It matters because a missing needle fragment ends up in a bag lining, and a brand that ships one has a product recall rather than a quality complaint.
What if a factory refuses a visit entirely?
A refusal for a first order is a significant signal, though genuine security or client-confidentiality reasons do exist. Offer a virtual audit with a fixed route as an alternative and judge the response; if both are refused, move to the next supplier.
How does a visit fit around sampling and production timing?
Visit after your sample is approved and before bulk starts, ideally at the pre-production meeting. Sampling takes 6 to 10 working days (12 to 15 for complex builds) and bulk runs 35 to 50 days, so a visit in the gap confirms that what you approved is what will be made.
What should the visit report contain?
Scores by area with the weighting you agreed, dated photo evidence, a ranked list of non-conformities, and a corrective action plan with an owner and a date for each item. Send it within 48 hours and ask for written acceptance before bulk production begins.
Get a Free Quote for Your Backpack Programme
Our production team works to MOQ 500 pcs, 6-10 working day sampling and 35-50 day bulk lead times, with final inspection to AQL 2.5 (ISO 2859-1). Send your brief and we will return a costed quotation in 24-48 hours.
Mens Backpack Guide Editorial Team — Sourcing guides and programme notes for the UK men's backpack market, coordinated by QUANZHOU JUNYUAN BAGS. Commercial figures such as MOQ 500 pcs, 6-10 day sampling and 35-50 day production reflect published programme terms.