Why Do Issues Still Slip Through Despite a Full Circle of Patrol Inspection? —— Five-Step Design Method for IPQC Patrol Inspection

When a customer conducts a factory audit and reaches the assembly station, they casually pick up the process patrol inspection records from the past month. The records are neatly organized: a row every two hours, all eight items marked with "√," and no signatures are missing. The auditor asks the on-site team leader, "What was the actual measurement of the crankshaft press-fit height during the patrol inspection at 3 PM yesterday?" The team leader cannot answer. A week later, a finished product is disassembled at the customer's site, and a missing shim is discovered—according to the patrol inspection records, the station was "inspected normally" that day.

Patrol inspection is the most cost-effective interception net in process inspection and is also the easiest to degrade into a mere signature exercise. It doesn't consume materials or wait for finished products; it just takes a few minutes to catch drifts before they turn into defects. However, if designed incorrectly, it becomes a formality: the records are full, but issues still slip through.

1. Why Patrol Inspection Becomes a Signature Exercise

Four common root causes:

• Points are set for convenience, not risk. The patrol inspector walks a convenient route, and the inspection points are naturally those on the way. The truly drift-prone and historically high-frequency processes are often left out.
• Items are copied from process cards without criteria. The form states "check if assembly is normal," but what "normal" means is undefined, making it always valid to mark "√."
• Frequency is uniform across the plant. Processes with a tool change cycle of two hours and those that remain unchanged for a year share the same "every two hours" frequency. The former is under-inspected, while the latter is over-inspected.
• Abnormalities are only recorded, not escalated. When a deviation is found, it is noted as "notified," but the form does not specify who will handle it, how long it will take, or whether the batch needs to be isolated.
• Only symbols are recorded, not data. Records filled with "√" symbols are useless for traceability and cannot prove whether the process was under control or not when issues arise.

To redesign patrol inspection as a genuine net, follow these five steps.

2. Step One: Set Points Based on Risk, Not Convenience

More inspection points do not necessarily mean better. Patrol inspection resources are limited, and redundant points only shorten the time spent at each, ultimately turning the inspection into a mere formality. The basis for setting points should be risk:

• Processes where special characteristics are marked in the drawings and control plan;
• Processes that drift over time: tool wear, mold temperature, thermal balance, glue quantity and formula, tension;
• Historically high-frequency nonconforming processes and those pointed out by customer complaints;
• High-risk windows after changes: new model changes, new employees, equipment maintenance, shift changes;
• Status of poka-yoke devices, process parameters, and material identification (to prevent material mix-ups).

Once the points are set, define the route: follow the material flow direction from start to finish without retracing steps. The route of a patrol inspection determines whether it can be consistently followed. A convoluted route will be abandoned within three months. Ensure that night shifts and shift changes are covered—these are the most common areas for missed inspections.

The output is a patrol inspection point list: point number, process, covered special characteristics or parameters, and responsible shift. This list corresponds to the process characteristics in the control plan, ensuring that "what to inspect" is no longer decided on the spot by the inspector.

3. Step Two: One Item, One Criterion, and the Criterion Must Be Clear

The sources of patrol inspection items are fivefold: process parameters and product characteristics in the control plan, special characteristics in the drawings and specifications, the status and effectiveness of poka-yoke devices, process parameters (temperature, pressure, time, torque), and the traceability of material and batch identification. Copy these actionable items into the patrol inspection form to avoid "inspecting what should not be inspected and missing what should be."

Criteria come in three forms, clearly specify which one:

• Numerical Range: Torque 45±3 N·m, mold temperature 80～90 ℃. Reserve a column for actual measurement values, not just ticks.
• Limit Samples: Appearance, burrs, shrink marks. Use physical samples to define boundaries, and ensure samples have a number and expiration date.
• Status Confirmation: Whether the poka-yoke device is powered and effective, whether the material rack is poka-yoke, and whether parameters have been altered.

If the criteria are unclear, the patrol inspection will revert to ticking boxes. A vague "check if normal" is more harmful than missing a point.

4. Step Three: Frequency Follows Drift Speed

A uniform frequency is the easiest and least effective design. Frequency is determined by two factors: how quickly the process drifts and how severe the consequences are after drift.

Cover at least once within the tool change cycle; inspect drift-prone parameters hourly; check appearance and assembly at the beginning and end of each shift; add a confirmatory patrol inspection after model changes, maintenance, and material changes. During the window period for new employees or recent process changes, increase both the points and frequency, and return to normal once stability is achieved.

The next step is the escalation line, which is often missing from patrol inspection forms: what actions are taken by whom and within what time frame when an abnormality is found. It is recommended to specify three levels:

1. Level One: Correctable on-site, correct and re-inspect during the shift, and record the re-inspection results.
2. Level Two: Affects continuous production, stop the line or isolate the batch produced during the shift, and notify the process owner.
3. Level Three: Involves special characteristics or has a risk of outflow, open a nonconformity, initiate corrective actions, and assess downstream batches and delivered products.

If the escalation line is not specific to positions and time limits, it is as good as non-existent.

5. Step Four: Leave Evidence That Can Self-Validate

The value of patrol inspection records lies not in "proving we inspected," but in answering "what was the process state at the time of the issue."

• Record actual measurement values and status, not symbols; in the abnormality column, write the phenomenon, the action taken, and the re-inspection results.
• Accumulate patrol inspection values by process and plot them on trend and control charts to supplement process capability analysis.
• Link records to batches, allowing retrieval by batch number and time period, so they can be presented when customers investigate.
• High-frequency issues should be fed back: if a patrol inspection abnormality item recurs three times within a month, it should be returned to the control plan, PFMEA, and work instruction to revise criteria or frequency, rather than continuing to tick boxes.

6. Step Five: Use Results to Review Patrol Inspection Effectiveness

The effectiveness of patrol inspection cannot be judged by the inspector's own statement but by result comparison. Three indicators are sufficient:

• Process Interception Rate: The proportion of issues intercepted within the process, compared to those found by subsequent inspections or customers. An increase in the former is a positive signal.
• Overlap: Whether the failure modes of batch nonconformities fall within the patrol inspection items and whether they are inspected at the required frequency. Items that slip through are gaps in the patrol inspection form.
• Data Completeness Rate: The percentage of rows in random checks where actual measurement values are fully filled in. This is a more honest indicator than "complete signature rate."

After each batch nonconformity, review the patrol inspection records of the process at the time—whether the frequency was followed, whether the criteria were covered, and whether the values are reasonable. Once these three checks are complete, the direction for revising the patrol inspection form becomes clear.

7. An Example

An automotive parts assembly line received a customer complaint: missing shims. Tracing back, it was found that a significant portion of the missing shims that were caught at the final inspection station could have been intercepted within the process. The process patrol inspection records showed all eight items marked with "√," including one that read "assembly completeness."

The rectification involved only four actions: restructure the patrol inspection items from eight to six, change "shim in place and poka-yoke device status" to a status confirmation item, inspect the first piece of each shift and every two hours; write the escalation line on the back of the patrol inspection form; add an inspection after model changes and before and after night shift changes; and summarize patrol inspection values weekly into trends. Three months later, the proportion of missing shims intercepted within the process increased from 12% to 68%, the final inspection interception rate dropped significantly, and there were zero recurrences on the customer side. The additional time spent by the inspector in each round was less than three minutes because they were inspecting six items with clear criteria, not eight undefined items.

Patrol inspection is not just filling out another form but using the least cost to intercept drifts before they become defects. Setting points based on risk, having clear criteria, frequency following drift, having escalation actions for abnormalities, and data feeding back into the system—these five steps ensure that patrol inspection records transform from signatures into evidence.

The value of patrol inspection lies not in how many lines are signed but in how many issues are intercepted.

Knowledge code: 11.1.1

Version: v20261001

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