What Causes Repeat Nonconformances on Projects?

Containment protects the project while the issue is being investigated. It can include stopping work, segregating material, increasing inspection, repairing affected work, or notifying downstream stakeholders. These are necessary controls, but they are temporary by design.

Problems begin when a containment action is logged as a completed CAPA. If the team repaired ten assemblies but did not determine why the process produced the error, assembly number eleven remains exposed. Closure should not be based on whether the original defect disappeared. It should be based on whether the process can now reliably prevent the defect from returning.

Human error is often a contributing factor. It is rarely a sufficient root cause. Statements such as operator error, lack of attention, or failure to follow procedure describe what happened, not why a capable person could make the same error in the operating system.

A defensible investigation tests the conditions around the individual: Was the current drawing available at point of use? Were acceptance criteria clear and measurable? Was the employee qualified for that exact process? Did the work sequence make inspection impractical? Did supervisory checks occur at the required stage? Was the required tool calibrated and suitable for the tolerance?

When the answer is simply retrain the employee, organizations frequently create a paper closure with no durable process change. Training is appropriate when a verified competency gap exists. It is not a substitute for correcting ambiguous instructions, inaccessible records, conflicting specifications, or inadequate supervision.

Many repeat nonconformances start upstream of the work face. A contract requirement may be understood by the owner but not incorporated into a purchase order. An engineering revision may be issued but not reflected in a supplier traveler, inspection plan, or field work package. A code requirement may appear in the quality manual but never become a measurable acceptance criterion.

This is especially common at interfaces between engineering, procurement, construction, and quality. Each group may believe another function owns the requirement. The result is predictable: people execute from different versions of the truth.

Corrective action must trace the requirement from its source through specifications, drawings, procedures, work instructions, inspection points, and final records. If the flowdown fails at any point, recurrence is likely even if the original work was repaired.

A quality management system can contain approved procedures, qualified personnel, and detailed inspection plans while the actual work process operates differently. Inspectors may arrive after concealment. Hold points may be waived informally. Production records may be completed at the end of a shift rather than as work occurs. Supplier surveillance may focus on document completeness while critical characteristics receive limited independent verification.

The issue is not whether a control exists. The issue is whether it operates consistently, at the right point, with evidence that can withstand audit or dispute.

For critical work, effective controls are built into execution. They identify who verifies the requirement, when verification occurs, what acceptance evidence is required, and who has authority to stop release when evidence is absent. Adding more final inspection may detect defects earlier, but it will not necessarily prevent them. The right balance depends on risk, process capability, and the consequence of escape.

A CAPA program weakens quickly when actions are assigned to departments rather than accountable individuals. Improve communication, review the process, and reinforce expectations are not corrective actions that can be verified. They are intentions.

Each action should identify an owner, a completion date, the affected process, the required deliverable, and the acceptance evidence for closure. For example, revise the receiving inspection procedure, train named personnel, update the inspection and test plan, and complete independent surveillance of the next three lots. That is specific enough to audit.

Ownership also needs to extend beyond the quality function. Quality can administer the nonconformance system, but engineering must own design clarity, procurement must own supplier requirement flowdown, operations must own execution discipline, and leadership must resolve resource or schedule pressures that make conformance difficult.

This is the most common reason a closed nonconformance becomes a repeat nonconformance. Teams verify that an action was completed, but not that it worked.

Completion evidence might show a procedure was revised or a training session occurred. Effectiveness evidence shows that the revised process produced conforming results over a defined period, under normal operating conditions. Depending on the risk, verification may include targeted audits, record sampling, first-article review, enhanced surveillance, trend analysis, or independent inspection of subsequent work.

The verification period should match the process. A high-volume fabrication issue may produce enough evidence in days. A low-frequency commissioning, maintenance, or supplier qualification process may require months or several cycles. Closing too early creates a false sense of control; holding every CAPA open indefinitely consumes resources. The appropriate standard is risk-based evidence , not administrative convenience.

Organizations often define nonconformance closure criteria after an issue has occurred. That invites inconsistency, particularly when schedule pressure is high. A better approach is to establish, before mobilization or production, which findings require formal root-cause analysis, who can approve closure, when independent verification is mandatory, and what evidence is required for recurring or safety-critical issues.

Trend analysis should also look beyond identical defect codes. A rise in incomplete records, missed inspection points, and unauthorized substitutions may represent one broader control failure: weak document control or inadequate field supervision. Categorizing nonconformances by process, supplier, work area, requirement source, and consequence makes patterns visible before they mature into repeat events.

Independent verification has particular value where internal production, schedule, or commercial incentives could influence closure decisions. An independent reviewer is not there to add paperwork. The role is to test whether the stated evidence supports the closure claim and whether the governing standard has actually been met.

A repeat nonconformance should change the level of scrutiny, not simply add another line to the register. Treat recurrence as a demand for better evidence, clearer ownership, and a corrective action that fixes the process before the process produces the next defect.