Corrective Action Report Example: A Filled CAR with Evidence

Separate five different decisions

A corrective action report should let another person follow the reasoning from a specific failure to a verified improvement. Problems arise when every response is entered under “action taken.” The report then hides whether the team controlled the immediate impact, fixed an affected item or changed the process that caused it.

• Containment: limit exposure while the issue is investigated.
• Correction: resolve the detected nonconformity or affected output.
• Cause analysis: establish why it happened using evidence.
• Corrective action: address the supported cause to prevent recurrence.
• Effectiveness evidence: demonstrate that the change achieves its intended result.

ISO 9001 Auditing Practices Group guidance distinguishes correction, cause analysis and corrective action and emphasises reviewing effectiveness. Its papers offer educational guidance rather than additional certification requirements. The worked example below is original and fictional, with dates and sample sizes chosen for that particular scenario.

Write a finding someone can investigate

Fictional example: Alder Packaging supplies printed cartons. Its internal procedure DSP-04 revision 3 requires the dispatch checker to verify the order revision against the released production order. During a 3 June 2026 audit, three of twelve sampled dispatch records showed revision B cartons sent against revision C orders. The auditor retains the order numbers, scan records and corresponding release documents.

The finding states that the dispatch revision-verification step was not effectively implemented for those three orders. This separates the requirement, the evidence and the conclusion. “Dispatch needs improvement” supplies none of those details. “The checker was careless” jumps to an unsupported cause before the investigation starts.

Give the issue a unique identifier and connect it to the originating audit. An internal audit report template can hold the finding and overall audit context while the linked CAR holds investigation and follow-up. Avoid maintaining two conflicting versions of the action status.

A filled corrective action report

The report links to the evidence rather than embedding every screenshot. Each attachment has a date, version and description so a future reviewer can identify what was tested. Customer details remain in the appropriately restricted system. The dates illustrate the sequence; they do not establish universal response periods.

Test the proposed cause

Alder first considers operator competence, stock identification and the dispatch screen. Interviews show that experienced checkers followed the same screen prompt. Replaying an affected order reproduces the stale revision. Reviewing stock labels finds the expected revision markings. These observations support the screen and workflow explanation more strongly than an assumption about individual attention.

The investigator also checks another carton family using the same software path and finds the same exposure. That result expands the action scope. A five-whys worksheet can help organise questions, but a neat chain of answers is not proof. Record the documents, observations or tests supporting each important link and retain unresolved alternatives until they have been examined.

The NCR forms pack provides templates for recording nonconformities and corrective actions. Use separate fields for the immediate response, supported cause and process change so a reviewer can assess their relationship.

Set an effectiveness test before closure

Decide in advance what result would support closure, who will review it and when enough relevant work will have occurred. In the example, the reviewer includes both shifts, checks actual released-order revisions and tests the mismatch control. The follow-up period contains real revision changes, making it more informative than a period when every order used an unchanged design.

A revised procedure and a successful software deployment demonstrate implementation. They do not, by themselves, establish effectiveness. Equally, zero complaints can hide an issue customers have not yet detected. Combine an appropriate result measure with evidence that the revised control operates as intended. Forty-eight dispatches is this fictional team's chosen review population, not an ISO sample-size rule.

If a mismatch recurs, keep or reopen the issue, assess the impact and revisit the cause and action. If insufficient relevant work has occurred, record why the effectiveness decision remains pending. A completed implementation date should never silently become a closure date.

Review the report with this checklist

• Identify the requirement, observed evidence, scope and precise failure.
• Record containment and correction with affected-output decisions.
• Support the cause with investigation evidence and check for similar exposure.
• Link each corrective action to the cause, an owner and a justified due date.
• Verify implementation using identifiable records and observed changes.
• Define the effectiveness method, review population, acceptance criteria and reviewer.
• Record the closure decision, outstanding risks and continuing monitoring.

For multiple findings, the audit templates kitoffers programme, report and tracking structures. Keep one traceable record per issue and review overdue or repeated findings together so individual close-outs do not hide a wider process problem.

The free audit evidence pack adds a worked evidence trail and editable log for connecting a finding to implementation and effectiveness records. Keep those references linked to the original CAR identifier.