Skip to content

A Reference Work for World History

What quality management system audit documents on site: part 2

By admin
The short answer

An audit report is only as strong as what the team documented while standing in the factory. Procedures can be read from a desk, but proof that incoming checks happen, that operators follow work instructions, that gauges are fit for use, and that bad product is actually segregated must be captured on site through records, observation, and lot tracing. Buyers who learn to read the evidence section can separate a rigorous visit from a polite walkthrough even when both reports use similar language. This article shows what complete on-site documentation contains and how to spot the gaps that undermine a verdict.

Key takeaways

  1. Judge every audit conclusion by the documented evidence behind it: which records, which observations, and which lots the team actually examined.
  2. Expect coverage of core record families such as incoming control, process monitoring, calibration, competence, and nonconforming-product handling, consistent with the system elements described in quality management standards. [2]
  3. Look for lot tracing that connects documents to real product, echoing the logic of acceptance checks that verify conforming output rather than paperwork alone. [1]
  4. Treat photographs as illustrations and dated records plus direct observation as the real evidence.
  5. Read scope, method, and exclusions before findings, because undocumented areas cannot support any conclusion.

Buyers often receive an audit report and jump straight to the findings page. That habit wastes half the document. What the audit team documented on site, which records they pulled, what they watched, and which lots they traced, determines how much weight every finding and every clean verdict deserves. Two reports with identical finding counts can support opposite sourcing decisions once the documentation behind them is compared.

The reason is straightforward. A management system audit builds its conclusions from observed evidence, not from the existence of a quality manual. Standards in the ISO 9000 family describe the system elements that should exist, from document control to monitoring of processes. [2] But whether those elements function at a given factory is something the auditor must establish молотком of records and observation during the visit. The report's documentation section is where that establishment work shows.

This article explains what competent on-site documentation looks like: the record families auditors should capture, how observation and document review reinforce each other, how lot tracing ties the system to real product, and how buyers should read the evidence section before acting on the verdict.

The record families that show a system in use

A functioning quality management system leaves a paper or digital trail across every shift, and the audit should sample each family of records rather than admiring the manual. Incoming-material records show whether purchased components are verified before use: delivery notes matched to purchase requirements, inspection entries with accept or reject decisions, and identification of rejected lots. Production records show whether defined controls were performed: process parameters logged at stated frequencies, in-process check sheets with real entries including occasional failures, and equipment settings recorded where they matter.

Measurement and calibration records come next. Auditors look for gauge lists, calibration status, recall arrangements, and evidence that out-of-tolerance equipment is handled rather than quietly returned to service. Competence records show who is qualified for which operation, with training and authorization entries that match the roster on the floor. Nonconforming-product and corrective-action records close the loop: segregation logs, disposition decisions, rework and re-inspection entries, and follow-up on causes. When any family is missing or contains only perfect entries with no rejections, that itself is a documented observation worth challenging.

Buyers reading the report should check that each family is mentioned with specifics: record titles, date ranges sampled, and what the sample showed. A sentence stating that incoming records for three named component lots were traced, with one rejection handled through a recorded disposition, carries real information. A sentence stating that documentation was generally adequate carries almost none. The discipline of naming records and dates is what separates documentation from decoration.

How observation confirms what records claim

Records can be completed after the fact, which is why competent audits pair every important record family with direct observation of the work it describes. The auditor watches the incoming check being performed, stands at the workstation while the operator follows the instruction, checks the gauge label and its calibration status on the equipment itself, and walks to the segregation area to see whether it exists, is identified, and actually contains the logged nonconforming product. Observation turns a claim in a file into a fact about the floor.

Illustration: How observation confirms what Decorative illustration for the section "How observation confirms what"; visual only, carries no data.

Interviews add the third leg. Asking operators to explain their task in their own words reveals whether training reached the people or only the training file. Operators who perform a controlled process describe the checkpoint, the frequency, and what they do when results drift; operators handed a form shortly before the visit describe the form. Good reports document these conversations in summary form, noting roles interviewed and whether explanations matched the written process. A quality management system audit worthy of the name spends most of its hours on this triangle of records, observation, and interview rather than in the meeting room.

Timing matters for observation quality. Announced audits let factories tidy up, so auditors should still ask to see normal evidence: earlier shifts' records, lots already shipped, rework from last month. Night-shift and weekend working deserve explicit attention where they exist, because controls that function at ten in the morning with management present may lapse at midnight. Reports should state which shifts and lines were observed so buyers can judge how representative the picture is.

Evidence typeWhat it provesWhat buyers should check
Incoming and purchasing recordsPurchased inputs verified against requirementsNamed lots traced, rejections and dispositions shown
Process and checkpoint recordsDefined controls performed at stated frequencyDate ranges sampled, entries across shifts, occasional failures visible
Calibration and gauge statusMeasurement results can be trustedGauge list, recall system, labels seen on equipment
Competence and training recordsRight people do critical operationsQualification entries matching the floor roster
Nonconforming product logsBad product is segregated and dispositionedSegregation area seen, rework with re-inspection entries
Traceability exerciseSystem connects to real shipped productLot traced forward and backward with record references
Management review and internal auditSystem is monitored by the factory itselfDated minutes and follow-up on internal findings

Lot tracing ties the system to real product

The strongest single piece of on-site documentation is a lot trace. Starting from a finished lot or a recent shipment, the auditor works backward to incoming materials and forward to delivery records, pulling the documents the system says should exist at each step. Where the chain holds, the factory demonstrates control of the lot, not just possession of procedures. Where it breaks, the breakage point names the exact process to fix, which is far more actionable than a general observation about document control.

Illustration: Lot tracing ties the system to real Decorative illustration for the section "Lot tracing ties the system to real"; visual only, carries no data.

This logic mirrors acceptance practice for finished equipment and goods: conformance is demonstrated on the actual output through defined checks before release. [1] In the audit context, the traced lot plays the role of the unit under test, and the records at each handover play the role of the test evidence. Buyers should therefore look for at least one fully described trace in the report, with lot identity, documents pulled, and the result stated plainly including any break.

A useful trace covers the buyer's product type wherever possible rather than whatever lot is most convenient. If the buyer's line was idle, the report should say so and trace the closest comparable production while flagging the substitution. Buyers commissioning audits can require this explicitly in the scope letter: name the product family to be traced and require the trace description in the report. Without that instruction, teams under time pressure trace the tidiest available lot, and the documentation looks complete while answering the wrong question.

Reading the evidence section like an auditor

Start with method and coverage, not findings. Note the visit dates, the areas and lines covered, the shifts observed, the records sampled with date ranges, the people interviewed by role, and the stated exclusions. This frame tells the reader what the report can legitimately claim. A finding of minor gaps across a broad, well-sampled scope is stronger assurance than a clean sheet from a narrow visit that avoided the night shift and the subcontracted process.

Then test each conclusion against its cited evidence. For every significant verdict, the report should let the reader answer three questions: which records support it, what was observed to confirm it, and which lot or order it was tested against. Where the chain is complete, the buyer can rely on the verdict within its stated scope. Where the chain references only a procedure title or a single photograph, the buyer should treat the verdict as provisional and request the missing evidence before placing consequential orders.

Finally, file the evidence where the next audit can build on it. Keep the record titles, the traced lot identities, and the exclusion list alongside the corrective-action log. The follow-up visit then starts from documented facts rather than from memory, and the factory understands that documentation standards rise rather than reset each year. Over two or three cycles this habit produces reports that get shorter and more factual, because both sides know exactly what evidence will be asked for.

Sources

  1. Wikipedia: Factory acceptance test —— Overview of factory acceptance testing performed before shipment or installation.
  2. Wikipedia: ISO 9000 —— Overview of the ISO 9000 family of quality management standards.

Frequently asked questions

How can buyers tell whether the documentation review was thorough?

Ask which records were sampled, over what period, and what the sample showed, including any gaps. A documented trace of a real order or lot through the system tells more than a list of procedure titles. If the report only says documents were reviewed without saying which and what they showed, request the missing detail.

The report includes many photographs. How much weight should they carry?

They are useful context but weak evidence on their own. Photographs show a moment; records across weeks show a routine. Give most weight to dated records, completed check sheets, and traceability exercises, and treat photographs as illustrations of conditions the auditor also verified in records.

Does a report with no findings mean the factory is fully controlled?

No. Documented findings describe gaps found within the visited scope on the visited dates, while excluded areas remain unknown either way. A report with zero findings and a narrow scope is less reassuring than a report with minor findings and a broad, well-evidenced scope. Read scope and findings together.

What if the report is missing evidence for a critical process?

Request a supplementary evidence round covering the missing area, ideally on site and during live production. State exactly which records and observations are needed. Do not let the gap slide until the next full audit, because production decisions taken in the meantime will rest on an incomplete picture.