Should Every Item on the Drawing Be Inspected? — A Five-Step Method for Selecting and Verifying IQC Inspection Items
In a certain automotive parts company, the IQC inspection item list grew from 27 items to 86 items over five years, with each item having a story behind it: customer complaints, returns, and claims. The inspection time per batch increased from 1.2 hours to 3.5 hours, and while inspectors began working overtime, missed inspections continued—last year, a batch was returned from the customer site due to a failure mode of bushing cracking after press-fitting, yet this issue was not even listed among the 86 items.
The problem does not lie in the diligence of the inspectors but in the source of the inspection items: they have always been added based on "what happened last time" rather than "how this batch might fail." As a result, the list keeps growing, but the risks are not adequately covered.
1. Inspection Items Come from Only Three Sources
Failure Modes and Process Capability. High-risk failures identified in the supplier's PFMEA, historical incoming material defects, and characteristics that have drifted on the SPC chart—these three sources should guide where the inspection items should be focused.
Customer and Regulatory Requirements. Special characteristics (SC/CC) on the drawing, customer special requirements (CSR), and regulatory limits (such as harmful substances and material composition)—these are mandatory inspection requirements.
Status Risks of Incoming Materials. Appearance, packaging protection, labeling and batch traceability, shelf life and storage conditions, transportation damage—these items are often not listed on the drawing but are frequent causes of production line stoppages.
Items that do not fall into these three categories are often "standard actions" copied from elsewhere. Typical characteristics include: inspected annually, never found defective, and unable to specify which failure mode they correspond to. Such items do not provide any interception value and only consume inspection time.
2. A Coverage Matrix to Clarify "What to Inspect"
The rows of the matrix are failure modes, and the columns are inspection methods and criteria. Example:
| Failure Mode (Source) | Inspection Item | Method/Tool | Sample and Frequency | Criteria |
|---|---|---|---|---|
| Insufficient plating thickness (supplier changed plating solution, 2 historical incidents) | Plating thickness | Magnetic film thickness gauge | 5 pieces/batch (batch ≥ 500) | Single piece ≥ 8μm, average ≥ 10μm |
| Bushing cracking after press-fitting (supplier mixed materials, 1 customer incident) | Material hardness + batch verification | Rockwell hardness tester + label verification | 3 pieces/batch | HRB 70～80, batch number consistent with COA |
| Sealing part aging failure (storage beyond shelf life) | Production date and packaging | Visual + label | Every batch | Remaining shelf life ≥ 6 months |
The criteria in the matrix must be numerical or clearly determinable, such as "good appearance" or "no obvious defects," which should not appear in the matrix. For attributes that are difficult to quantify (appearance, assembly interference), boundary samples, photos, or limit samples should be provided to ensure consistent judgment by different inspectors. Even the best-designed projects cannot prevent issues if the criteria are not clear.
The value of this matrix lies in its bidirectional traceability: each item can be traced back to a specific failure mode, and each failure mode can be traced to at least one inspection item. Failure modes that cannot be traced to an inspection item should not be discarded but clearly noted as "controlled by the supplier's process, no IQC inspection item set." The risk has not disappeared; it has merely been transferred. It must be clear who is managing it, otherwise it will be forgotten.
3. Five-Step Method: Backward Tracing from Failure Modes to Projects
Step One: List Failure Modes. Summarize incoming material defects from the past 12 to 24 months, including those detected by IQC, found on the production line, and returned by customers. Categorize and deduplicate these by failure mode, noting the frequency and consequences. Data should come from at least four sources: IQC inspection records and nonconforming product handling forms, daily production line defect reports, customer complaints and return reports, and supplier change notifications (material, process, and origin changes). The defects found on the production line and at the customer site are the easiest to overlook, yet they are the gaps in coverage. Supplier change notifications are often forgotten but are often the starting point for "sudden issues in old projects."
Step Two: Match Each Failure Mode with an Inspection Item. For each failure mode, assign an operational inspection item and criteria that the inspector can use to make a judgment. If this is not possible, such as for internal defects or material internal stress, mark it as "not inspectable with current methods" and proceed to Step Three.
Step Three: Change the Inspection Method for Non-Inspectable Items. Switch to supplier process control (require process data with each batch, implement poka-yoke in critical processes), perform incoming verification tests (periodic destructive sampling), or tie supplier performance (relax inspection only after continuous qualification). Avoid using a vague "strengthen appearance inspection" as a solution, as this is ineffective.
Step Four: Reverse Simplification. Review each existing inspection item: if it cannot be traced to a failure mode and has no customer or regulatory basis, delete it or reduce its frequency to quarterly verification. Allocate the saved inspection time to SC/CC characteristics and items that have previously caused issues. Inspection time is limited, and the matrix serves as the allocation table.
Step Five: Set Up Feedback and Review. If a defect appears on the production line or at the customer site, first check if it is covered in the matrix: if not, add the item; if it is covered but not intercepted, review the inspection method and criteria. Conduct a full review of the matrix every six months and update it when changing suppliers, materials, or processes.
4. Three Common Misuses
More Items Equal Stricter Control. 86 items may seem thorough, but in reality, they dilute the inspectors' time, leaving critical items under-inspected. The interception capability depends on the strength of the critical items, not the total number of items.
Using Inspection Items to Replace Process Control. If the drawing tolerances are loose and the supplier's plating capability has not been validated with Cpk, relying on full IQC inspection is costly and yields unstable results. The correct approach is to return to DFMEA/PFMEA and the supplier's process capability, rather than endlessly adding items at the incoming inspection stage.
Matrix Completion and Wall Hanging. If new defects are not fed back into the matrix and standards are not updated when changing suppliers, the matrix will become disconnected from reality within three months, leading to the old cycle of "adding an item wherever there is a complaint"—this is precisely why the inspection item list expanded to 86 items.
The number of inspection items does not reflect the control strength; a bidirectionally traceable coverage matrix does. The goal of inspection is to ensure that each item corresponds to a specific failure mode.
The number of inspection items does not reflect control strength; a bidirectionally traceable coverage matrix does.
Knowledge code: 9.2.1
Version: v20261013
Author: QTank QTank is dedicated to providing systematic knowledge, methodologies, and practical tools for quality management professionals, helping companies continuously improve their quality capabilities.