A good sample shows that a factory can make an acceptable garment. It does not show that every shift can repeat the result across an order of 10,000 pieces. Bulk work brings new fabric lots and changed settings. Different operators must repeat the result while the shipment date draws near.
QA and QC help turn that approved result into daily work. The factory needs clear rules and sound methods. It also needs checks that reveal when output no longer meets the standard. It also needs a way to act before the same fault spreads.
For a Bangladesh RMG factory, this is a practical management task. A folder of procedures matters only when people can use it. An inspection result matters only when the team knows which work it covers and what to do next.
Two factories, two ways of finding trouble
Picture two hypothetical factories on the same industrial road in Dhaka. Both have approved samples and experienced sewing teams. Their methods for controlling bulk production differ.
Factory A relies mainly on inspection near the end. Cutting, sewing and finishing move ahead before the main quality review. If a fault repeats, a large quantity can reach the repair area at once. The factory may still ship good garments, but it must recover the work within the remaining time.
Factory B agrees key checks before production. It places them where staff can find faults sooner. It confirms materials and approves the first output. When staff raise a concern, the team responds. Final inspection remains part of the system.
Factory B has more opportunities to limit the spread of a fault. That does not prove its results are better in every case. Look at the work itself and the defect trend. Customer feedback and lasting fixes provide further evidence. A written system alone cannot establish that performance.
What QA and QC mean on the factory floor
Quality control, or QC, covers methods that help daily work meet quality requirements. Inspection and testing are part of QC, alongside controls applied during production. QC is not confined to checking finished garments.
Quality assurance, or QA, is planned work. It builds confidence that the work will meet its requirements. It includes how the factory plans, maintains and checks its quality system. ASQ explains this QA and QC distinction. It also notes that their tasks overlap.
For a seam, QC might check stitch density, appearance and the specified performance. QA might define how to approve and teach the method. It can also define how to control changed settings. A sound system needs both the result and evidence that the method can be repeated.
Factory question | Example QC activity | Supporting QA activity |
|---|---|---|
Is this fabric suitable for the order? | Inspect and test the received lot | Define approval rules and manage supplier performance |
Does the seam meet the specification? | Check the seam and record the result | Approve the method, resources and training |
Can this bundle move forward? | Confirm its inspection and hold status | Define release authority and the handoff process |
Did a correction work? | Check later output for the fault | Review the cause, action and evidence of lasting effect |
Is the order ready for release? | Perform the agreed final checks | Ensure the plan, records and decision process are valid |
Job titles vary between factories. A QA inspector may check products. A QC manager may also improve the quality system. Define the work and authority clearly rather than relying on the title alone.
Build a quality management system around the work
A quality management system, or QMS, connects requirements, processes, people and evidence. ISO's introduction to a QMS describes linked processes and clear roles. The system must fit the business. For an apparel factory, the system should follow how an order actually moves.
An internal quality audit can test whether the defined controls are being used and supported by traceable evidence. Keep its criteria and sampling frame with the findings and follow-up record.
The five areas below provide a practical starting structure. They do not form a full checklist for certification. Buyers may use different audit models.
1. Policy and goals that guide decisions
A quality policy should help people resolve a real choice. An output target may conflict with an agreed hold rule. Staff need to know what to do and who can decide. A broad promise of “best quality” gives little help in that moment.
Translate the policy into useful goals. A line may seek fewer repeated seam faults. Stores may focus on keeping unapproved material out of cutting. State the measure, baseline, review date and responsible person. A target without a defined counting method can create arguments instead of improvement.
Management must also provide the time and resources needed to meet those goals. A request to cut rework loses force if managers also remove the first-piece check.
2. Roles and authority that survive a busy shift
Assign responsibility for material approval, process checks, holds, repairs and release. Identify a backup when the usual person is absent. The night shift needs a workable answer. It should not depend on a favour from another team.
Quality staff need a clear route to raise concerns and prevent suspect work from moving. Production staff still own how the work is made. Inspection alone cannot make a process reliable. Each team must take care of quality in its own work.
A role table names who performs each check. It also names who reviews exceptions and approves release. Keep those decisions separate where the risk or buyer's procedure requires it. Repair and release are separate decisions. The person who repairs a piece may lack release authority.
3. Processes that explain the handoffs
Map the order from its requirements to dispatch. Include material approval, cutting, sewing and finishing. At each handoff, define the inputs and expected output. Name the record that shows their status.
For example, cutting needs the approved pattern revision and usable fabric. Sewing needs correctly identified panels and approved accessories. Finishing needs the right care label, packing instruction and measurement basis. A gap can remain when each team thinks another team has settled these details.
Work instructions should cover the features that matter to the task. A close-up photo or approved sample may help. A short method with pictures can also be easier to use than a long paragraph. Name an owner for revised instructions. That person must ensure that each affected workstation gets the change.
4. Communication and demonstrated skill
The person doing the work must be able to understand its instructions. Use language, pictures and examples suited to the team. Ask the operator to demonstrate the method and explain what to do when a problem appears.
A training attendance sheet proves attendance. It does not prove that the person can perform the operation to the required standard. Supervised practice and checked output provide stronger evidence of competence.
Inspectors need the same support. They need sound methods and clear defect samples or guides. If two inspectors disagree on the same garment, resolve the cause. Their counts cannot fairly rank lines while their judgment differs.
5. Feedback that changes the process
Defect records should lead to a response with an owner and a review date. The same fault may recur after an operator was instructed. A repeated note does not show a lasting fix. The team needs to know whether it found the cause and whether the fix worked.
Review comparable results over time. A lower defect count may reflect less output or a different style mix. Record how much was checked, at which stage and under what conditions. Managers can then judge what changed: the work or the report.
Customer complaints belong in this review too. A fault that escaped factory checks may reveal a gap. Review the inspection method, handling controls and test plan. Treat the complaint as evidence to investigate, not merely a claim to close.
Convert sample approval into a bulk quality plan
An approved sample does not answer every production question. The team also needs the current specification and its limits. Approved materials, tests and packing instructions belong in the plan. Resolve any conflict between those sources before using them in bulk work.
ISO 10005's public scope covers quality plans. It offers guidance on how to create, review, use and revise them. It is guidance, not a separate set of requirements. A factory can use the planning idea whether or not it seeks ISO 9001 certification.
For a specific style, a practical plan might contain the following fields:
Plan field | Example of the question it answers |
|---|---|
Requirement and revision | Which approved specification applies? |
Characteristic | What dimension, seam, shade or label feature matters? |
Method | How will the feature be measured or checked? |
Timing and coverage | When will checks occur, and what work will they cover? |
Acceptance rule | What result meets the requirement? |
Record | Where will the result and piece identity be kept? |
Reaction | What must happen if the result fails? |
Authority | Who can hold, review and release the work? |
Choose checks based on the product and process risk. A new seam or wash may need closer checks. A stable repeat task may need less attention. Increased attention does not mean inventing a universal hourly sample size. Use the agreed buyer method and a reasoned internal plan.
Control materials and changes before they spread
Material status should be visible before use. An arrival record shows that fabric reached the factory. Approval is a separate step. Stores and production need a clear status for each material. Is it awaiting review, accepted or on hold?
Link fabric rolls, trims and relevant approvals to the order. Keep shade groups or lots separate where the order needs this. Their identity must survive cutting and assembly. An inspection result needs a clear link to the goods checked. Without that link, it has limited value.
Changes deserve the same discipline. A changed thread or pattern may affect output. So may a new supplier or a different pressing setting. Name who reviews the change and when it takes effect. Define the trial or approval it needs.
Remove old instructions from use. Keep the history that explains how earlier work was made. A printed copy pinned beside a machine can remain wrong long after a central file has been updated. Revision control therefore needs a floor check, not only a file-management rule.
The fabric receiving checklist provides a more detailed starting point for incoming-material controls.
Make measurement results dependable
Two people can use the same tape and still measure a garment differently. They may choose different points or apply different tension. They may also measure at different stages of relaxation. The method needs to specify the relevant conditions, not simply name the instrument.
Check equipment condition and suitability for the tolerance being judged. A calibration record is useful. It cannot correct a poor method of measuring the garment. NIST's traceability guidance explains a key limit. A traceable measurement may still be unfit for a specific use.
For a borderline dimension, follow the agreed decision rule and method. Do not round a failure into acceptance because shipment is urgent. If the requirement is unclear, ask the authorised party to clarify it in writing.
Product traceability and measurement traceability also mean different things. One links a garment to its materials and production history. The other concerns the reference basis of a measurement result. A factory may need both, but a bundle ticket does not establish instrument accuracy.
Early checks reduce the quantity exposed to a fault
Consider a hypothetical operation producing 120 pieces per hour. A setup error begins immediately after a change. If the first effective check occurs three hours later, 360 pieces may need review. A check after twenty minutes could limit the review to forty pieces.
These figures describe exposure, not a claim that every piece is defective. Use the evidence to set the hold scope. Consider the fault and the last confirmed good output. The example shows why the timing of detection matters.
The response must follow the signal. Identify and separate suspect work. Correct the cause and confirm good output before the agreed restart. An early finding has little value if the affected bundles keep moving to the next stage.
Coats' sewing guidance illustrates why diagnosis matters. Similar stitch faults may have different causes. These can include needles, machine timing, thread control or material handling. A general instruction to “be careful” does not resolve those different mechanisms.
The Kanban handoff guide covers the cutting-to-sewing flow. It explains how to limit work in process. Its quantity rule must sit alongside a separate quality status. A card that authorises replenishment does not automatically authorise release of suspect goods.
Separate right first time from successful repair
Right first time means meeting the agreed requirement on the first pass. The piece needs no unplanned repair or rework. Catching a fault early is valuable, but the affected piece is still a first-pass failure.
Suppose 1,000 garments receive their first check. Nine hundred pass, while one hundred require repair. Eighty repaired pieces later pass. In this example, twenty remain rejected.
First-pass success = 900 ÷ 1,000 × 100 = 90%.
Final accepted share = (900 + 80) ÷ 1,000 × 100 = 98%.
Both measures are useful. Reporting only 98% hides the effort used to recover eighty pieces. Reporting only 90% hides the final disposition. Keep the first result and link later repair checks to it.
If the one hundred faulty pieces contain 120 faults, DHU is 12, while the faulty-piece share is 10%. The DHU guide explains how to use those measures without confusing their denominators.
Correct the garment and improve the system
Repairing a seam corrects that garment. To stop a repeat seam fault, act on its confirmed cause. These are related jobs, but completing the first does not prove the second is finished.
For a recurring loss, a DMAIC project charter helps state the problem, measurement basis and project boundary before the team selects a solution.
ASQ's Plan-Do-Check-Act model provides a useful sequence for testing change. For example, a team first defines the fault and likely cause. It then tests a changed setting or method and checks later output. The evidence guides whether to adopt or adjust the change.
An effectiveness review needs enough relevant production to be meaningful. One good piece after an adjustment may support restart. It may not prove that the fault is gone. Heat, shift length or changing material could bring it back.
Record what was changed and what remained constant. Include the results that did not improve. A failed trial can still narrow the search. Keep an honest record and control the suspect product.
Statistical process control can help detect unusual changes in a suitable process measure. It does not keep the process stable by itself. Control limits differ from the buyer's specification limits. The team still needs the correct rules for product acceptance.
Use records that a colleague can follow
A useful record allows someone else to reconstruct a decision. It identifies the order, style, lot, date, process stage and specification used. It shows what was checked, the result, any hold and who approved the next step.
Paper records can do this well when they are clear and controlled. Digital records can be easier to retrieve. They still need access controls and a revision history. Define how to handle missing or corrected entries. Software does not remove the need for a sound definition.
The manufacturing analytics guide connects these records with decisions. Start with the question the team must answer. Which bundles are affected? Did the correction work? Add automation where it improves that answer.
Keep standards and buyer requirements current
Certification, buyer approval and shipment acceptance have different scopes. A quality-system certificate has a defined scope. It does not replace the product specification or decide an order's status.
ISO states that ISO 9001:2026 replaces the 2015 edition. A certified factory should confirm the applicable transition arrangements with its certification body. For an audit, confirm the agreed edition and scope. Use the licensed requirements where needed.
The practical structure in this article does not claim to reproduce those requirements. Its purpose is to make daily quality work easier to perform, trace and improve. Follow one live order through the five areas. Then close the gaps that the evidence reveals.
Part 2 covers buyer audit preparation, AQL, traceability and corrective action. A strong audit trail begins with reliable work on an ordinary production day.
Discussion0
Loading the discussion…