Garment Factory Evaluation: Verify Technical Capability Before Placement

Evaluate the factory against the garment it must make

A polished sample display cannot answer every question about a garment factory. The pieces may use different fabrics or come from a separate sample room. A specialist who made them may not handle the new order. Technical evaluation needs a closer link between the proposed product and the evidence shown.

The question needs a clear focus. Can this site make the required construction from these materials? Can it repeat the result and detect a problem before it spreads? A factory may be strong in one product group while needing development work for another.

This guide uses an invented adult woven blouse to explain the evaluation. It has a narrow front placket, a fused collar, curved hems and long visible seams. These features provide practical test questions. The example quantities and tolerances are fictional. Each buyer must set the rules that apply to its own product.

The purpose is a technical decision before placement. Price, available production time and workplace conditions also matter. Each needs its own evidence. A good sewing trial does not approve those other areas.

Turn the product into a short technical risk list

Start with the current design, material description and sample requirements. Identify the few features most likely to challenge the proposed process. This keeps the visit focused and helps the factory prepare useful evidence.

For the blouse, the evaluator might select four questions. Can the factory keep the placket width consistent? Can it fuse the collar without changing the required handle? Can it control the curved hem without twisting? Can it keep the long seams smooth before and after the intended care process?

Those are questions to investigate, not claims that the fabric will fail. The factory may demonstrate a sound method for each. It may also identify a design or material issue that the buyer has missed.

Keep the list short enough to follow in detail. Four well-chosen demonstrations can reveal more than twenty tick boxes with no records behind them. Add other questions when the product or intended use requires them.

Distinguish an essential feature from a preference. If the placket width is essential to the design, the factory must show how it will control it. If a construction detail can change, define who may approve the alternative. The evaluator should not redesign the garment by accepting an unrecorded shortcut.

Know what each form of evidence can establish

A technical file can contain claims, records, observations and test results. They carry different weight because they answer different questions.

A machine list states what equipment is reported to exist. A live demonstration shows how selected equipment performs during that observation. A previous test report describes the named specimen and method. Repeated results from the relevant process provide a broader view. Check which materials and period those records cover.

Use plain status terms such as “claimed”, “observed”, “supported by records” and “not yet demonstrated”. This makes uncertainty visible. A missing demonstration leaves a question open. It does not prove that the factory lacks the skill.

For example, a supplier may have made many lightweight woven garments. It may still need a trial in the proposed blouse fabric. That is a gap to close. A recent trial in a similar cloth may help plan the work. The actual material still needs its own evidence.

ISO 9001 concerns quality management systems. A relevant certificate can be part of the factory's evidence. The blouse measurements and seam results still need to be checked. Check what a document covers and use product-specific evidence for product-specific questions.

Follow one component through the proposed route

A factory tour is more useful when it follows a real item. Select an identified component and trace its instructions, materials, operations and checks. For the blouse, follow the collar. Its result depends on the fabric and interlining as well as fusing, sewing and pressing.

Begin with the pattern version. Ask how the cutting team knows it is current and how the matching pieces are identified. Then follow the fabric and interlining to the proposed fusing step. Check that the trial uses the intended materials, or that any substitution is stated.

Ask the operator or technician to explain the chosen process. Which settings came from supplier guidance? Which were confirmed by the factory's trial? How are the finished components checked? The purpose is to see whether decisions can be explained and repeated.

Continue to sewing and finishing. Look at how matching parts stay together. Check how the team keeps a rejected component away from good work. Trace the final collar back to the trial record. A result that cannot be connected to its materials and settings is hard to reproduce.

If an operation takes place elsewhere, record the actual site and the handoff. An off-site process can be valid, but its technical evidence belongs in the route. An impressive main factory cannot demonstrate an outside process merely by describing it.

Observe a relevant operation, not just the machine

Ask for a demonstration that resembles the proposed work. Material and ply arrangement can affect the result. So can seam direction, thread, needle and attachments. A straight seam on a firm scrap does not resolve a curved hem in a soft blouse fabric.

Record the conditions that make the demonstration meaningful. Identify the fabric and component, the person doing the work and the equipment used. Note any preparation or special handling that would also be needed in production.

Watch the start, middle and end of the operation. A neat middle section can hide distorted entry or exit points. Examine both faces and the inside of the seam where the construction allows. Watch for repeated corrections. Ask whether these are part of the planned method or extra work needed for this sample.

The demonstration should also reveal how the method is taught. Ask another trained operator to explain the instruction. You can also review how the factory confirms that training has worked. One skilled sample maker may produce a sound result. The factory must also show how that method will work in production.

Do not confuse this with counting machines or calculating line capacity. Technical suitability and available production time are separate decisions. The evaluation should identify the process needed. The production team must then confirm that the agreed route can be scheduled.

Check whether the team can diagnose a defect

Technical competence includes the ability to investigate a poor result. Ask the team to explain a relevant problem. What evidence showed its cause? Which check confirmed that the correction worked?

Seam puckering is a useful example because the same visible symptom can have several causes. Coats' technical guidance discusses thread tension, fabric structure, differences in material movement and changes during care. It shows why blaming the thread alone can miss the source of the problem.

During evaluation, ask what the team would examine first on the blouse's long seam. A reasoned answer should connect the proposed check to a possible cause. “Use more pressing” is incomplete if the seam returns to the same condition after handling or care.

Where a trial change is appropriate, retain the original and changed pieces. Record what changed and what stayed the same. The team may change the needle, thread, tension and fabric together. An improved result can then be useful, but it cannot show which change caused it.

Avoid pushing the team toward a predetermined answer. The material or design may need attention as well as the sewing method. A good evaluator asks what the evidence shows. A plausible explanation still needs a test before it becomes a finding.

Confirm that measurements can support the decision

Before interpreting the trial, agree what is measured and under which conditions. Name the point, garment state, unit and measurement method. Compare widths at the same agreed stage. Pressing and care may change the dimension.

Check the measuring equipment used for the task. Its readable divisions, condition and measurement uncertainty must suit the required decision. A calibration label by itself does not answer those questions.

NIST's policy on metrological traceability describes a documented chain of calibrations. That chain connects a measurement result to its reference. It also stresses that traceability alone does not guarantee fitness for purpose. The uncertainty must be suitable for the measurement need.

For the blouse, ask the factory to show its method. Watch how it finds the finished placket edges and handles the garment. Two people may use different points or apply different tension. Their readings can then differ even with the same ruler. Resolve the method before calling the difference a production defect.

Where a result lies near a limit, apply the agreed decision rule and measurement basis. Do not add extra decimal places to make an uncertain reading appear decisive. Keep the original reading and seek a valid repeat or further assessment when needed.

A useful average can still hide a failed piece

Suppose the fictional brief sets a finished placket width of 30.0 mm with a tolerance of ±1.0 mm. For this arithmetic example, the limits are inclusive at 29.0 mm and 31.0 mm. Assume the measurement method and decision rule have already been agreed.

Eight trial garments give the following readings:

Mean width = 241.0 ÷ 8 = 30.125 mm

The mean is within the example limits, but garment 6 is outside them. Reporting only the mean would hide that result. Seven of the eight observed garments meet the stated width condition.

Eight pieces do not establish long-term process capability. They provide evidence about this trial and a reason to investigate the wide result. Keep the individual readings, material identities and order in which the pieces were made. A gradual change and one isolated handling error may need different follow-up work.

Examine the complete care and performance evidence

An acceptable fresh sample may change after the intended care route. Plan the relevant checks before the visit or trial. This helps the factory prepare a useful test instead of an improvised demonstration.

ISO 6330:2021 provides domestic washing and drying procedures for textile tests. Its public scope makes clear that a complete test includes both washing and drying. It also notes that changes in the procedure can affect the result.

For the blouse, a report that says only “washed” leaves important questions open. Ask which agreed route was used. Confirm the number of cycles and the checks made afterwards. Match the conditions to the product brief. A procedure name alone does not create a buyer's acceptance limit.

Inspect the relevant dimensions, seam appearance and component behaviour on identified specimens. Keep the before-and-after states clear. Record any pressing required by the method. If the method does not call for it, extra pressing could hide the observed change.

An illustrative length check shows why the baseline matters. Suppose an agreed marked length is 80.0 cm before care and 78.8 cm afterwards:

Length reduction = 80.0 − 78.8 = 1.2 cm

Reduction relative to the starting length = 1.2 ÷ 80.0 × 100 = 1.5%

This is a calculation, not a pass decision or a full test procedure. Whether the result is acceptable depends on the agreed requirement and method. It also says nothing by itself about a twisted hem or a collar that has lost its intended shape.

ApparelDiary's fabric shrinkage guide develops the dimensional-change topic further. During evaluation, check how the team uses this evidence. Can it obtain the right results and explain what they mean for the product?

Select seam tests by what they actually measure

A seam test report should name the method, specimen and result. Avoid treating every report headed “seam strength” as interchangeable.

For example, ISO 13935-2:2026 concerns maximum force to seam rupture using a grab method. Force is applied across the sewn seam. Its stated scope includes woven fabrics, with defined limits and exclusions. The public page identifies it as the replacement for the withdrawn 2014 edition.

That scope does not establish a universal minimum force for all garments. It also does not turn a result from a different fabric or construction into evidence for the blouse. Confirm the relevant edition and full method with the laboratory. Apply the acceptance criteria agreed for the product.

The standard's public abstract helps choose a test; it is not the complete procedure. Owning a tensile tester is only part of the task. A useful result needs the right method, trained staff and suitable specimens.

Review laboratory competence and the actual report

Some factories test in-house; others use an outside laboratory. Either route needs clear responsibility for sample identity, test instructions and results.

ISO/IEC 17025:2017 addresses laboratory competence, impartiality and consistent operation. It provides a different kind of evidence from a general quality-management certificate.

The International Accreditation Service explains how a laboratory's scope defines its accredited activities. It does not automatically cover every test the laboratory offers. Where accredited testing is required, look for the relevant method and activity in that scope.

Then read the report itself. Match the sample code, fabric, construction and date to the trial. Check the required conditions against the report. Look for any stated departure from the method. A valid report for an old specimen may provide useful background. It does not approve the new product.

Ask who reviews the result and what happens when it does not meet the brief. Sending a report to the buyer is not the same as understanding it. The factory should identify the affected work and propose a sound next step. It must also preserve the original finding.

Keep first-pass results separate from repaired results

A factory may achieve an acceptable final sample through extra repair. Record that repair work. Readers should be able to see which pieces passed the first check.

Suppose a fictional trial contains 12 garments. Nine meet the agreed inspection conditions on their first inspection. Three need repair, and all three pass the later reinspection:

First-pass accepted share = 9 ÷ 12 × 100 = 75%

Final accepted share after repair = 12 ÷ 12 × 100 = 100%

Both figures describe the same trial at different stages. The second does not erase the repair work. Record what was repaired, why it was needed and whether the change can be built into the method before a larger run.

These are observed trial proportions. They are not a forecast of the factory's future defect rate or a basis for accepting a shipment. The trial is too limited for that conclusion, and its conditions may differ from later production.

Also distinguish a defect count from the number of garments needing repair. One garment may contain several faults. Define the unit before comparing records. Count each garment once in the original trial, even if it later needs repair.

Use past records to test whether the system works

Follow a previous technical issue through the factory's records. Remove customer details where needed. Choose a relevant construction or material problem. The useful evidence is the link from observation to action and then to a checked result.

For instance, a record may show that a collar changed shape after care. Can the team identify the affected material combination and the trial used to investigate it? Can it show which instruction changed and how later work was checked? A final photograph alone leaves most of that process unexplained.

Check the dates and sample identities as you move through the record. A later test on an unrelated fabric does not close the earlier finding. A changed interlining may resolve one construction while leaving another untested. Keep the scope of the correction clear.

If the factory shares a performance percentage, request its numerator, denominator and period. Find out whether it covers all inspected work or selected successful orders. Compare definitions before drawing a conclusion. Past records can show how the system behaves over time. The current blouse trial answers the narrower question about the proposed product.

Close findings with evidence of the corrected method

An evaluation should leave each open issue with a clear next action. State the observed problem, the affected product feature and the evidence needed to close it. Assign a responsible person and an agreed review date.

For the wide placket, a weak action would be “train operators”. A stronger record would show the measured variation and proposed method change. It would also define a repeat trial to assess the result. Training may be part of that work, but it is not proof that the result has improved.

Ask to see the revised instruction or tool where relevant. Then review new pieces made under the revised conditions. Keep them linked to the earlier trial so the improvement can be assessed. Do not replace the original measurements with the new ones.

Check transfer to the intended production route. One specialist may make the corrected sample. If bulk work uses a different method, that route still needs to be checked. The follow-up should show how staff will learn and check the process at the production site.

Issue a product-specific decision

The conclusion should state its limits as clearly as its result. Name the blouse and the route found suitable. List the issues that must close before placement. An unexplained factory score cannot convey that scope.

The result may support placement within the scope assessed. It may require a hold for more evidence, or show that the proposed route is unsuitable. Set those decision rules before the final review. Keep required performance results visible. Attractive premises or friendly communication cannot cancel a failed requirement.

Record the product version, site, materials and processes assessed. Identify the people and documents involved, the trial results and any work that was outside the review. Keep capacity approval, commercial terms and other required assessments as separate decisions.

ApparelDiary's quality assurance and quality control guide links preventive systems with product checks. Look for that link during the visit. Ask the factory to show a result and explain the method that can repeat it.

Revisit the decision when the production basis changes

An evaluation has a defined scope, not an unlimited life. A new fabric, interlining, sewing route or subcontract site can change the evidence needed. Even when the style name stays the same, the technical basis may no longer match the approved trial.

Define which changes require review and who can authorise them. Keep the approved sample, relevant files and test records accessible. ApparelDiary's PLM guide describes how those product records connect across development.

The strongest factory evaluation ends with a usable decision file. It records what the factory showed and what remains uncertain. It also sets the checks for the next stage. The merchandiser can then make a placement decision, and the factory knows what the work requires.

References

• ISO: quality management systems and ISO 9001.

• Coats: technical causes of seam puckering.

• NIST: policy on metrological traceability and fitness for purpose.

• ISO 6330:2021: domestic washing and drying procedures.

• ISO 13935-2:2026: seam maximum force using the grab method.

• ISO/IEC 17025:2017: competence of testing and calibration laboratories.

• IAS: laboratory accreditation, competence and scope.