How Do Inspectors Compare Goods with Approved Samples?

Inspectors compare randomly selected production units side by side with the identified approved sample, using the current specifications and drawings. They check identity, appearance, measurements, function, and packaging within the agreed scope, then record the differences. Written tolerances and defect definitions determine whether a checked characteristic conforms; the agreed sampling plan determines how the findings affect the lot result.

At UTS, we use the approved sample as a physical reference, not as a substitute for written requirements. Our reports present actual findings, not corrective-action or release recommendations. Shipment decisions remain with the buyer and supplier.

Before mass production, our sample evaluation checks sample units against drawings, specifications, and design documents. That establishes what the reference represents before it is used to compare production goods. [1]

When the Comparison Takes Place

These are the production checkpoints described in our inspection services, not percentages of the order that will be inspected:

What Makes an Approved Sample Usable?

An approved sample needs an identifiable approval record and a defined scope. It may show the accepted colour and finish without defining every dimension, internal material, or test requirement. Before using it, confirm:

• The sample ID, version, approval date, and location. For example, “GS-01, version 2” is an illustrative identification format, not a product requirement.
• Who approved it and which characteristics or limits that approval covers.
• Whether it is sealed, signed, photographed, or otherwise identifiable and protected from substitution or damage.
• The specification, drawing revision, bill of materials, packaging artwork, and checklist that apply alongside it.

If the sample is missing, altered, or inconsistent with the confirmed documents, our team records the limitation. Checks supported by the confirmed documents can continue within the agreed scope, but characteristics that depend on an unavailable physical reference cannot be presented as a completed side-by-side comparison. Photos alone do not establish texture, fit, weight, or hidden construction.

How Our Team Prepares the Comparison

In our initial-production inspection workflow, we review the purchase order, specification sheet, approved sample, and quality requirements, then customise the checklist for the product. These are the working materials behind the comparison, rather than a generic checklist applied without reference to the order. [2]

Our sample-evaluation process also checks dimensions against engineering drawings and documents identified defects with photographs. This gives the comparison a written and measured basis instead of relying on visual resemblance alone. [1]

To prepare the inspection, send our team:

• The purchase order and SKU list, with quantities, models, colours, sizes, and destination versions.
• The approved sample identity, approval record, and the physical reference where required.
• Current drawing revisions, measurement points, units, tolerances, and measurement methods.
• Defect classifications, the sampling plan, and acceptance criteria for the inspection stage.
• Packaging, labelling, shipping-mark, and barcode or QR-code files by SKU and carton type.
• Test instructions, equipment needs, test quantities, and any separately agreed laboratory-testing scope.

If artwork, material, components, or packaging changed after approval, identify the new revision and confirm which characteristics of the existing sample still apply. Different files may have different revision numbers; what matters is that each is approved for the same order and product configuration.

How the Side-by-Side Comparison Works

Confirm the Lot and Select Units at Random

Our team identifies the inspection lot, available quantities, carton list, and SKU coverage. Units are selected at random from the available production goods under the agreed plan, rather than relying on factory-prepared display pieces. The approved sample is the reference and is not included in the production sample count.

Acceptance sampling uses information from a random sample to support a lot acceptance or rejection decision. It does not inspect every item and is not primarily a method for estimating the exact quality of a lot. [5]

ISO 2859-1 defines AQL-indexed sampling schemes for lot-by-lot inspection by attributes. The agreed standard, edition, inspection level, lot size, and sampling rules must be identified in the inspection plan; an AQL value alone is not a sample size or an acceptance count. [6]

The sample quantity for a functional or destructive test may differ from the quantity used for visual checks. The record should therefore state coverage for each check, not imply that every selected unit received every possible test.

Compare Product Identity

A matching colour does not make the wrong SKU acceptable. For example, SKU-B units found in a carton marked SKU-A are recorded with the affected carton identity even if they look similar to the approved sample.

Compare Appearance Under Consistent Conditions

Place the sample and selected unit side by side under the same lighting, viewing distance, and orientation. Depending on the product, the checklist covers colour, pattern alignment, surface finish, stitching or welds, print position, logo clarity, and cleanliness.

Apply the agreed defect definitions consistently. A scratch, stain, open seam, or colour deviation is judged against the specified size, location, count, and viewing conditions, not a new standard invented during the inspection. Where an acceptable range is defined with boundary samples, identify which boundary applies to the characteristic being checked.

Measure Against Written Tolerances

Dimensions, weight, thickness, and fit are checked with the agreed methods and suitable equipment. Instrument suitability matters as well as calibration: NIST explains that metrological traceability alone does not establish fitness for purpose, because measurement uncertainty must suit the measurement need. [7]

In a hypothetical example, a drawing specifies 200.0 mm ±1.0 mm, giving an allowed range of 199.0–201.0 mm. A reading of 201.8 mm is outside that range even if the unit looks like the sample. This is an illustrative tolerance check, not a universal decision rule; any agreed rule for measurement uncertainty also applies.

Test Function Against the Agreed Method

For the units allocated to each functional check, use the agreed method, sequence, conditions, and pass criteria. Depending on the product, checks may cover assembly fit, opening and closing, switches, power behaviour, motion, sound, or light output.

Record the test conditions, cycles or duration where applicable, and observed result. A difference from the reference is assessed against the approved functional requirements; resemblance to a sample cannot replace a defined test.

Some composition, restricted-substance, safety, and durability questions require laboratory methods rather than visual or basic on-site checks. Where included in the agreed scope, our team coordinates separate testing with the sample identified to the order. ISO/IEC 17025 sets requirements for the competence of testing and calibration laboratories; citing it does not establish the accreditation of a particular laboratory or make an on-site inspection a laboratory test. [8]

Check Packaging, Labels, and Codes by SKU

Compare product labels, instructions, retail boxes, inner packaging, master cartons, and shipping marks with the approved files. Check the SKU assortment and count in each inspected pack or carton, as well as any destination-market markings included in the scope.

Where barcodes or QR codes are in scope, our team scans the inspected codes using the stated equipment and conditions and matches the decoded data to the approved SKU or carton data. The required scanning success rate is 100% for those inspected codes. This does not mean that every code in the shipment has been scanned or that every later device will read every code.

A successful scan is not the same as a print-quality grade. Where formal grading is specified, ISO/IEC 15416 covers linear barcode print quality and ISO/IEC 15415 covers two-dimensional symbols. That verification requires an agreed method and suitable verification equipment; it is not implied by an ordinary scanning check. [9][10]

How Differences Are Recorded and Assessed

For an observed mismatch, the record identifies:

• The characteristic, location, and approved reference used.
• The actual condition, with measurements, test results, or photographic evidence where applicable.
• The affected sample units and their SKU, carton, or pallet identity.
• The defect classification under the agreed definitions.
• Any missing reference, unavailable equipment, or uncompleted check that limits the comparison.

UTS service pages describe zero acceptance for critical defects, with commonly used AQL settings of 2.5 for major defects and 4.0 for minor defects. The order-specific plan controls the applicable settings. Zero acceptance means that a critical defect found in the inspected sample is unacceptable; it does not prove that the uninspected goods contain no critical defects. [11]

Keep two judgments separate: whether an observed characteristic is nonconforming, and whether the sample findings meet the lot acceptance criteria. AQL does not turn an out-of-tolerance measurement into a conforming one. Sample size and the plan's acceptance and rejection numbers determine the statistical lot result, while any separately agreed mandatory requirements still apply. [5][6]

Our inspection report presents the actual scope, references, observations, measurements, test results, defects, and checks that could not be completed. It does not include corrective-action, improvement, or release recommendations. Preparation guidance in this article is general information, not report content.

A suspected material change or outdated file is not automatically a proven cause. If the available evidence does not establish why a difference occurred, the report records the observed difference and the limits of the evidence without inventing an explanation.

What a Match Does Not Prove

• One matching unit does not establish lot acceptance. The result depends on the agreed selection and sampling plan.
• A newer factory sample is not automatically approved. Buyer approval and the applicable documents must identify the reference in force.
• A visual match does not establish hidden composition or durability. Those characteristics require the records or tests specified for them.
• A passed sample inspection is not a zero-defect or legal-compliance guarantee. It supports a lot decision within the agreed scope, with sampling risk; it does not verify every unit. [5]

Choose the Inspection Stage for the Comparison Needed

The same reference can be used at different production stages, but the available goods and inspection purpose change:

• Initial production inspection at 5–10% produced compares early output with the approved sample and specifications. [2]
• During-production inspection at 30–50% finished and packed checks consistency as production progresses. [3]
• Final random inspection when 100% is produced and at least 80% is packed checks the finished goods and available packaging before shipment. [4]

Choose the stages according to the product, supplier history, changes since approval, and agreed inspection scope. An earlier comparison can reveal a mismatch while production is still underway; a final comparison cannot retrospectively check how earlier units were made.

The practical answer is to identify the reference, select production units under the agreed plan, compare each in-scope characteristic under consistent conditions, and document the evidence. The sample shows the approved physical reference; written requirements define conformity, and the sampling plan governs the lot result.

Sources

1. UTS — Sample Evaluation: reference documents, dimensional checks, and defect documentation.
2. UTS — Initial Production Inspection: production stage and pre-inspection review.
3. UTS — During Production Inspection: mid-production timing and comparison scope.
4. UTS — Final Random Inspection: final-stage readiness and checks.
5. NIST/SEMATECH — What Is Acceptance Sampling?: random sampling and lot disposition.
6. ISO 2859-1: AQL-indexed sampling schemes for inspection by attributes.
7. NIST — Metrological Traceability: traceability, uncertainty, and fitness for purpose.
8. ISO — ISO/IEC 17025: testing and calibration laboratory competence.
9. ISO/IEC 15416:2025: linear barcode print-quality assessment.
10. ISO/IEC 15415:2024: two-dimensional barcode print-quality assessment.
11. UTS — Final Random Inspection, AQL Classification: critical, major, and minor defect settings.