QE Capability Advancement (26) | Safety and Regulatory Characteristics Management: A Closed Loop from Identification to Monitoring

1. An Expired List for Two Years

A certain electronics manufacturing company underwent a customer audit. The auditor did not start by reviewing the drawings but requested something else: the revision records of the product safety and regulatory requirements list. The quality engineer retrieved the list, which was dated two years ago and had not been updated since. The auditor then asked three questions: When was the last time the mandatory standards of the target market were benchmarked? Are the hazardous substance limits listed in the list consistent with the latest third-party reports provided by suppliers? How long would it take to identify all affected finished products if a batch were found to use materials not listed in the list?

The on-site team could not provide complete answers to these questions. The audit conclusion itself was not severe—list not updated regularly, incomplete regulatory evidence chain, and lack of actual traceability drills. However, the underlying issues were significant: the company's management of safety and regulatory characteristics had stagnated at the stage of "copying a list from the past," lacking both identification mechanisms and transmission and monitoring mechanisms. The cost of failure for ordinary characteristics is rework and cost, while the cost of failure for safety and regulatory characteristics is personal injury, recalls, and legal liability. Managing both types of characteristics with the same intensity is equivalent to betting on the risk that "no one will come to check."

2. Why Safety and Regulatory Characteristics Must Be Managed Separately

The difference is not in terms of "importance" but in three hard constraints.

Consequences are Non-Negotiable. For general characteristics, deviations can be handled through concessions, repairs, or downgraded use. However, the consequences of safety characteristic failures are irreversible, and regulatory limits are mandatory baselines, leaving no room for concessions. This means that control methods cannot rely on "post-event selection" but must focus on "pre-event prevention."

Responsibility is Traceable. After an incident, the three key questions for accountability are: Was it identified? Was control implemented? Can it be traced? The focus is not on whether it was intentional. These three questions depend on records. Without evidence of identification, control, and traceability, the responsibility naturally falls on the company.

Different Evidence Requirements. General characteristics require a pass/fail determination—tested, passed, and released. Safety and regulatory characteristics require a complete evidence chain: identification basis (which regulation, which customer requirement), control execution records (poka-yoke validation, SPC data), and traceability records (batch and identification). The retention period must comply with regulatory or customer requirements, which are generally longer than the company's internal retention periods for ordinary records.

These three constraints translate into a closed loop in management: identification → classification → transmission → monitoring → traceability, with an additional change review. Any break in this loop renders all previous work ineffective. The most common breakpoints in most companies are not in identification but in transmission and traceability—lists are well-maintained, control plans are written, but there are no corresponding clauses or record fields at the workstations, making it impossible to identify batches when an incident occurs.

3. Practical Steps and Quantitative Criteria

Step One: Identify from Four Sources, Each with a Written Conclusion

Safety and regulatory characteristics cannot be determined by "expert judgment" alone; they must be identified from four sources simultaneously:

1. Regulatory and Standard Sources: Map the mandatory standards of the target market to product characteristics, including safety regulations (GB/UL/CE), hazardous substances (RoHS/REACH), and industry-specific regulations.
2. Customer Requirements: Special characteristic symbols on customer drawings, customer-specific requirements (CSR) documents, and regulatory lists provided by customers.
3. Design Analysis Outputs: Failure modes in DFMEA with severity S ≥ 9 are automatically considered candidates for safety characteristics; those with S ≥ 7 and involving regulations are also included.
4. Historical and External Information: Internal accident and recall records, industry recall announcements, escalated complaints, and significant material changes from suppliers.

Criteria: Identification must be documented, with written conclusions (hit/miss and reasons) for each of the four sources. It is not acceptable to simply write "no issues found after evaluation." If a source is marked as "miss," the basis for this decision must be clearly stated; otherwise, it is considered not identified.

Step Two: Classify by Failure Consequences, Bind to Control Intensity

• Safety/Regulatory Class (commonly marked as CC or customer-defined symbols): Failure leads to personal injury, legal violations, or recalls. The target capability is Cpk/Ppk ≥ 1.67 (or as specified by the customer, some require 2.0); control methods include 100% effective control (poka-yoke preferred) + real-time SPC monitoring.
• Key Class (commonly marked as SC): Significantly impacts function, assembly, or subsequent processes. The target is Cpk ≥ 1.33, with sampling inspection + SPC.
• Important and General Class: Affects secondary functions or appearance, typically without independent capability thresholds.

The value of classification lies in differentiating three aspects: capability targets, monitoring intensity, and abnormal response speed. If any of these three aspects are missing, classification is reduced to just a symbol.

Step Three: Transmit Characteristics Through Six Documents

The transmission chain is a common audit hotspot, and the chain is:

Characteristics List (the sole authoritative source) → DFMEA/PFMEA → Control Plan → Work Instruction → Inspection Record Form → Traceability Mark.

Criteria: For any safety characteristic, the symbol, name, and number must be consistent across these six documents; the "Characteristics Level" column in the control plan must be filled; the work instruction must include corresponding clauses and record fields—the absence of record fields is equivalent to non-transmission, which is the most common hidden breakpoint. Any addition, deletion, or modification in the characteristics list triggers updates to the PFMEA and control plan, revision of on-site documents, and training confirmation, with a suggested closed-loop timeline of 15 working days.

Step Four: Monitoring Must Have "Fail-Proof" Evidence

• Poka-Yoke Devices: Verify the first piece at the start of each shift and conduct human-induced failure tests regularly (suggested weekly) to ensure the device is effective. The verification records must include at least one real interception result; simply writing "verified as qualified" does not count as evidence.
• SPC: For safety characteristics, collect subgroups every 2 hours or every N pieces, and use control charts to identify special causes. Note: a stable process that is not in compliance is still an unacceptable state.
• Regulatory Characteristics (e.g., material composition): Monitor using third-party reports from suppliers + COA for each batch/ incoming verification. The report's validity period must correspond to the batch and should not cover multiple batches across different publication cycles.
• Capability Revalidation: Conduct a capability revalidation quarterly or per batch, and immediately upon material, mold, equipment, or supplier changes.

Step Five: Traceability Must Be "Drilled" to Count

Criteria: Conduct at least one bi-directional traceability drill every six months—select a random finished product and trace back to the material batch, equipment, personnel, and parameters; select a random batch of raw materials and identify all affected finished products. The target time for identification is suggested to be ≤1 hour (or as required by the customer/regulation). Record retention periods must comply with regulatory or customer requirements; in the automotive industry, this is commonly 15 years or the product life + 1 year, while in medical and food industries, it follows respective regulations and should not be uniformly treated as 3 years internally.

4. Common Misconceptions

Misconception One: Treating Regulatory Characteristics as Customer Requirements. Identifying only based on customer drawings and ignoring the mandatory regulations of the target market. These two are not entirely overlapping—customer drawings may not cover additional limits for export markets, and regulations are the baseline, which should not be decided based on whether the customer "mentioned it."

Misconception Two: Building Lists Without Updating. No update cycle, no responsible person, and no change triggers, making the list a historical document. It is essential to establish two mechanisms: "at least one full benchmarking per year + immediate triggers for changes in target market regulations."

Misconception Three: Treating Symbol Copying as Transmission. The drawing has CC, the control plan states "inspect according to the drawing," but the work instruction lacks corresponding clauses—identification has not been converted into executable control methods, which is a typical case of "formal transmission."

Misconception Four: Using Full Inspection as a Substitute for Capability Proof and Poka-Yoke. The miss rate of full inspection increases with batch size and fatigue, and it does not provide process capability evidence. Safety and regulatory characteristics typically require 100% effective control methods (poka-yoke preferred) + capability data, with full inspection as a supplement.

Misconception Five: Only Conducting Forward Traceability. Only checking records by batch and not performing reverse identification, making it impossible to identify affected products when an incident occurs. Traceability plans that have not been drilled should be considered invalid.

5. Self-Check List

• Is the last full benchmarking date of the safety/regulatory characteristics list within one year of the target market's mandatory standards? Are the update triggers and responsible persons clearly defined?
• Does the identification cover the four sources: regulatory standards, customer requirements, DFMEA (S≥9), and historical accidents and recalls, with written conclusions for each?
• For any safety characteristic, can consistent symbols and numbers be found in the characteristics list, PFMEA, control plan, work instruction, and inspection record form, and does the work instruction have a corresponding record field?
• Do poka-yoke devices have records of human-induced failure tests, and do these records show at least one real interception?
• Have reverse traceability drills been conducted? Is the time to identify all affected finished products from any batch of raw materials within the target time as required by the customer/regulation?

The effort in managing safety characteristics lies not in identification but in maintaining an unbroken closed loop.

Knowledge code: 8.2.2

Version: v20261006

Author: QTank QTank is dedicated to providing systematic professional knowledge, methodologies, and practical tools for quality management practitioners, helping companies continuously improve their quality capabilities.