Supplier Defect Evidence Pack: Photos, Measurements and Traceability

A useful supplier defect report lets another competent person identify the item, understand the observation, reproduce the check and see which requirement was not met. A folder of close-up photos without sample identity, scale or acceptance criteria cannot do that reliably.

Build the evidence pack before debating root cause or remedy. It should preserve the sample, context, method, actual result, requirement and affected scope without overstating what the sample proves about the whole batch.

For an Australian importer, the pack may start with an offshore inspection before shipment or with goods received in Australia. In either case it must preserve enough context for the buyer, supplier and qualified reviewer to examine the same evidence across locations. The pack cannot by itself establish Australian product-safety, regulatory or legal compliance.

Start with a case ID and containment state

Create a unique case ID as soon as the issue is found. Link every photo, measurement, sample and message to it.

Record:

• product, SKU and controlled revision;
• PO, supplier, production site and shipment/batch identifiers;
• receiving location and discovery date;
• quantity received, quantity checked and sampling basis;
• immediate stock status and containment action;
• person who found the issue; and
• decision owner.

If potentially affected stock could be used or shipped, preserve the hold in your existing quality process until an authorised disposition is recorded.

Define the requirement before the defect

Do not describe an item as defective without identifying the relevant requirement or approved reference. That may be:

• a drawing dimension and revision;
• material or performance specification;
• approved golden sample;
• packaging artwork or construction revision;
• inspection instruction and acceptance criterion;
• PO requirement; or
• product-specific regulatory or safety requirement confirmed by a qualified owner.

Quote or extract only the relevant requirement and retain the controlled source. If the baseline is ambiguous, record “requirement unresolved” and hold the conclusion.

Photograph in a repeatable sequence

Use a sequence that moves from identity to detail.

1. Context image

Show the carton, pallet, lot, work area or sample group with identifying labels where appropriate. Protect personal or sensitive information.

2. Whole-item image

Show the complete product and orientation. Include the sample ID.

3. Location image

Show where the suspected defect sits relative to recognisable features.

4. Detail image

Capture the condition in focus under useful lighting. Take more than one angle if reflection, depth or surface texture matters.

5. Scale or measurement image

Include an appropriate scale, gauge or measurement display when it helps. The image supplements the measurement record; it does not replace method and instrument details.

6. Comparison image

Where valid, show the affected sample beside an approved reference or conforming sample with both identities visible.

Do not edit pixels in a way that changes the evidence. Cropping, rotation or exposure correction should be applied conservatively, with originals retained. If annotations are added, save them as a derivative and keep the original file.

Give every sample and image an identity

A file name such as IMG_8742.jpg loses context quickly. Use a controlled scheme:

CASE-SAMPLE-VIEW-SEQUENCE.ext

For example: NCR-0264-S07-DETAIL-03.jpg.

The evidence index should map the file to:

The checksum helps detect silent file replacement. It does not prove the image is truthful or the sample representative.

Record measurements as data, not pictures

For every material measurement, record:

• characteristic and location;
• unit;
• specified requirement and tolerance;
• actual result;
• method or work instruction;
• instrument ID and relevant status;
• environmental or setup condition where material;
• operator and timestamp;
• repeat readings if the approved method requires them; and
• sample identity.

ISO educational guidance notes that monitoring and measuring resources should support methods needed for valid results. It does not tell you which instrument, resolution, setup or tolerance is correct for a particular product. Those decisions belong to qualified technical owners.

Distinguish observation, measurement and interpretation

Keep three fields separate:

This separation makes later review easier. It also prevents a tentative cause—such as “mould wear”—from being recorded as an observed fact.

Describe the affected scope honestly

The checked sample supports conclusions about that sample under that method. Broader conclusions require a defensible sampling basis and consistent batch traceability.

Record the denominator: “4 affected among 32 checked from cartons 2, 7, 11 and 14” is more useful than “many defective”. Do not call the batch-wide defect rate unless the sampling plan and result support that statement.

Where incoming inspection found the issue, preserve the incoming quality inspection record and link it rather than retyping results.

Build the evidence index

The table is an index, not the evidence itself. Store files in controlled access with original metadata preserved where possible.

Send the supplier a bounded problem statement

The initial notice should state:

1. identified product, order and batch;
2. requirement and revision;
3. observation and actual result;
4. checked quantity and affected samples;
5. current containment state;
6. evidence-pack link or controlled transfer;
7. questions that need response; and
8. the next decision date or operational constraint.

Avoid prescribing the root cause before the supplier investigates. Ask the supplier to confirm immediate containment and trace the affected process, then use the supplier corrective-action plan for root cause, corrective action and effectiveness evidence.

Quality checks before release of the pack

• Every file maps to a case and sample.
• The requirement is identified and current.
• Measurements include method, unit and instrument identity.
• Originals are retained; annotations are derivatives.
• Photos are focused, correctly oriented and sufficiently contextual.
• Checked and affected quantities are explicit.
• Suspected cause is labelled as hypothesis, not fact.
• Personal, confidential or unrelated information is removed or protected.
• The pack does not make a batch-wide, safety or legal conclusion beyond the evidence.

Common weak evidence

• A single close-up with no product identity or scale.
• A ruler held at an unknown angle as the only dimensional record.
• Screenshots copied through chat until metadata and resolution are lost.
• A defect description that does not name the requirement.
• Photos from mixed batches stored in one unindexed folder.
• The supplier's explanation inserted into the observation field.
• An “approved sample” that has no controlled identity or revision.
• Edited images with no retained original.

The handoff to corrective action

The evidence pack is ready when a competent reviewer can trace the item, requirement, observation, method, result and affected scope without relying on the original observer's memory.

It does not prove root cause or decide the remedy. It gives containment, supplier response and corrective-action work a reliable starting point—and makes later commercial discussions less dependent on contradictory photos and recollections.

Sources

• ISO APG: Evidence collection
• ISO APG: Monitoring and measuring resources
• ISO guidance on documented information