Chapter 13: Step 5 — Effectiveness Verification and CAR Closure

A corrective action is not closed until you've proven it works. This is where many manufacturers fail. They implement a fix, declare victory, and move on—only to see the same problem resurface six months later.

Three Acceptable Verification Methods

1. Data-Based Verification (Strongest)

Collect quantitative data showing the problem is resolved. This requires:

• Clear acceptance criteria (e.g., "zero defects in 100 consecutive units," or "process capability Cpk ≥ 1.33")
• Pre-correction baseline (e.g., "defect rate was 2.3% before correction")
• Post-correction data collection (minimum 30 samples, preferably 100)
• Statistical analysis: run chart, SPC control chart, or trend line showing stability

Example: > "Corrective action: Increase inspection frequency on injection molding final check. Verification: Collected dimensional data on 100 consecutive parts following implementation on 2026-01-22. 100% conformed to specification. Cpk = 1.78. No nonconformances detected. Data attached as Appendix A."

2. Audit-Based Verification (Medium strength)

Schedule an internal audit specifically to verify the corrective action was implemented and is working.

• Audit scope: "Verify tool change interval on Molding Line 3 is now 500 parts; verify work instruction WI-2050 Rev D is being followed; verify assembly torque check is documented on 10 random work orders."
• Evidence: Audit checklist, auditor notes, photos of updated procedure signage
• Finding: "Audit conducted 2026-02-15 by [Auditor]. All three corrective action elements observed in place. Assembly operators demonstrated understanding of new torque step. No discrepancies noted."

3. Time-Based Verification (Weakest, use sparingly)

If the nonconformance was a one-time event with no obvious preventive control, you can verify by demonstrating no recurrence over a defined period. Use only for minor CARs or where root cause was a training gap that cannot be easily quantified.

Example: > "Nonconformance: Shipping label applied incorrectly on 5 units. Root cause: New staff not trained on label orientation. Corrective action: Provide training and observe next 20 shipments. Verification: Period of 2026-01-22 through 2026-02-28 (38 days). No recurrence of labeling errors detected in shipments or customer complaints. CAR closed 2026-02-28."

Common Closure Mistakes That Auditors Flag

Mistake 1: Closing Before Data Collection "We implemented the corrective action on 2026-01-22. It's working. Let's close it."

Auditor response: "Where is your verification data?" You don't have any yet because you closed before collecting samples. Reset the verification start date and recollect data.

Mistake 2: Accepting 'Training Completed' as Effectiveness Evidence Corrective action: "Train all staff on new procedure." Verification: "Training completed 2026-01-20. 12 employees attended."

Auditor question: "How do you know the training was effective? Did employees understand? Are they following the new procedure?" Training completion is an implementation record, not effectiveness proof. You need observation, testing, or defect-free performance data.

Mistake 3: Verification Data Doesn't Match the Root Cause Root cause: "Measurement system error (gauge calibration lapsed)." Corrective action: "Recalibrate gauge." Verification: "Produced 50 parts; all inspected okay."

Wait—that doesn't verify the gauge works. Your verification should be: "Gauge recalibrated by certified metrology service on 2026-01-23 (receipt attached). Recalibration interval reset to 6 months. Baseline parts inspected with recalibrated gauge and reference standards; all measurements matched within acceptable tolerance."