Cover: AI-generated conceptual illustration for WEFTSIGNAL, not documentary photography.
Two questions that look like one
A garment fails a wash test and someone asks why inspection did not catch it. The question contains a category error. Inspection examined finished units and found nothing wrong with them; testing would have characterised how a material behaves under a stated condition. The first looks at the thing in front of it, the second answers a question about a property. Confusing them leads to buying one and expecting the other, which is how a programme ends up with thorough inspection records and no evidence about the material at all.
Inspection examines a unit
Inspection is a judgement about an object: this garment, at this moment, against this specification. It sees workmanship, symmetry, seam integrity, trims, labelling, packing and measurement. Its evidence is the unit, and its record is a count of what was found. It is good at detecting what can be seen, felt and measured on a finished piece, and it is limited to what is present in the sample examined.
Testing characterises a material or a property
Testing takes a specimen and subjects it to a defined procedure, producing a result that describes behaviour rather than appearance. Dimensional change after care, colour fastness, tensile and tear strength, pilling, abrasion, fibre composition — these are properties, and a property is established by a method. CottonWorks, the textile education platform published by Cotton Incorporated, puts the point directly in its quality testing material: it is critical that fibres, yarns, fabrics and garments adhere to the particular specifications for each product.
The unit of evidence is different
The distinction that matters operationally is what each activity produces evidence about. Inspection produces evidence about the pieces examined and, through sampling, an estimate about the lot they came from. Testing produces evidence about the specimen tested, the method it was tested by and, if the sampling was designed for it, the material population that specimen represents. A test result is not a statement about a shipment, and an inspection result is not a statement about a fabric.
Controlled conditions are not optional in testing
A test result is only reproducible if the conditions around it are controlled. ISO 139:2005 defines the characteristics and use of a standard atmosphere for conditioning, for determining the physical and mechanical properties of textiles, and a standard alternative atmosphere that may be agreed between parties. CottonWorks makes the same point operationally, noting that reliable and consistent testing depends on proper product sampling and conditioning. Inspection has no equivalent requirement, because it is not producing a measurement of a material property.
Where relying on one for the other breaks down
Two failures recur. The first is inspecting a material problem: a fabric that will shrink beyond specification may pass every visual and dimensional check on the day, because the property that will fail it emerges only after care. No amount of additional inspection finds it, and adding inspectors to the line is the wrong remedy. The second is testing a workmanship problem: no laboratory result will detect a misaligned placket or a skipped stitch, because those are properties of how a unit was made rather than of the material it was made from.
A third, subtler failure is using a test result as a lot certificate. A dimensional-change result from one specimen does not license a claim that the shipment will behave the same way, any more than a passing inspection licenses a claim about the fabric.
The distinction is not purely theoretical. For example, a documented pre-shipment case in which a collar failed a manual stretch-and-release check, was corrected, and passed the same check before release — with the account stating explicitly that the check is not an instrument-based elasticity or durability test.
Why a laboratory’s competence is a separate subject
Because a test result is only as good as the laboratory producing it, laboratories are assessed on their own terms. ISO/IEC 17025:2017 sets out general requirements for the competence of testing and calibration laboratories, and it is the reference buyers commonly look for when a result will be relied on between parties. Ordering a test from an unassessed laboratory is not the same act as ordering it from an assessed one, even when the method named on the report is identical.
An illustrative example: the same complaint, two causes
Take, illustratively, a returned order with two problems. Some units have a twisted side seam; others have a shade that moved after the customer washed the garment. The first is a workmanship defect and belongs to inspection, which can detect and classify it on a finished piece. The second is a material behaviour that no inspection of the dry garment would have found, and it belongs to testing — a fastness method, run under controlled conditions and agreed with the customer. One complaint letter contains both, and no single activity answers it.
Deciding which instrument a question needs
Each of these belongs to a different activity:
- Appearance and workmanship: inspection, on the unit, against a classified defect list.
- Material behaviour: testing, by a named method, on a specimen drawn by a stated plan.
- Specification adherence: testing, with the specification stating requirement and method together.
- Conditioning and atmosphere: part of the testing protocol, stated in the report rather than assumed.
- Laboratory standing: the framework the laboratory works under, recorded with the result.
- Boundary of the claim: what the result does and does not license anyone to say about the shipment.
Using both, and saying which is which
Neither instrument substitutes for the other, and a programme that uses both still fails if the paperwork does not distinguish them. The testing requirements belong in the fabric specification checklist and the inspection requirements in an inspection plan. Where the two are conflated, the result is a file of certificates that answers no question anyone actually asked.
Source and further reading
Sources consulted 16 October 2026: ISO — ISO/IEC 17025:2017, General requirements for the competence of testing and calibration laboratories; ISO — ISO 139:2005, Textiles: Standard atmospheres for conditioning and testing; CottonWorks — Quality Testing. Standards are cited by designation and official catalogue page only. The decision list and the worked example here are WEFTSIGNAL editorial material, and none of the organisations named endorses WEFTSIGNAL. WEFTSIGNAL and NOVITA are separate businesses. However, the operator of WEFTSIGNAL also works on NOVITA’s digital marketing. This relationship is disclosed for transparency. Continue exploring quality control coverage and our editorial standards.
