Inspection Frequency in Control Plans: Still Guessing? —— A Five-Step Method for Designing Inspection Methods and Frequencies

1. One Line of Text, Two Outcomes

In the control plan of a certain automotive parts company, both SC (special characteristic) and general characteristics are inspected at the same frequency: a sampling inspection of 5 pieces every 2 hours. During a customer audit, the auditor asked, "Why every 2 hours? Why 5 pieces? What is the basis for this?" The quality department could not answer and received a nonconformity. What made the quality manager even more uncomfortable was another issue: last month, a critical dimension exceeded the tolerance, and the defect was only discovered by the customer. The "sampling inspection of 5 pieces every 2 hours" in the CP had been strictly followed—while the frequency itself was not wrong, it was far from sufficient for SC characteristics.

Inspection frequency is the easiest column in the control plan to be filled in by guesswork. The person writing it cannot explain the basis, and the person reviewing it can easily spot the issue. If the frequency is set too loosely, it cannot prevent nonconforming products; if set too tightly, it can exhaust the inspectors and skyrocket costs. Frequency is not about "how strict you want to be," but rather a balance between risk and cost.

This column also affects two areas: it is a frequent nonconformity during audits and serves as the daily execution guideline for inspectors. Clarifying this column means simultaneously safeguarding both the audit and the production floor.

2. Understand Three Variables Before Discussing Frequency

Inspection frequency is determined by three variables, all of which are essential.

Before starting, distinguish between two things: inspection method answers "how to inspect," while inspection frequency answers "how often to inspect." If the method is wrong, even the most frequent inspections will be in vain; if the frequency is too low, even the correct method will miss risks. Both must be decided together, and the key to frequency design lies in the following three variables.

The first is the consequence of failure. The more critical the characteristic, the greater the cost of missing a nonconforming product, and the more frequent the inspections should be. SC characteristics (safety, legal, fit and function) are the baseline, and their frequency can only be increased, not relaxed; CC characteristics (key characteristics) are next; general characteristics prioritize cost.

The second is process stability. Process stability is not based on feelings but on data: if Cpk is greater than 1.33 and the control chart is consistently under control, the frequency can be relaxed; for new processes, recent mold changes, recent parameter adjustments, or Cpk less than 1, the frequency must be strict.

The third is the cost of discovery. The later a nonconforming product is discovered, the more expensive it becomes: it costs 1 yuan to detect in the current process, 10 yuan in the next process, and 100 yuan if it reaches the customer. The design of inspection frequency essentially answers "if we don't inspect, where will the nonconforming product flow to?"

This is why processes closer to the customer often have higher inspection frequencies—not because they are more prone to issues, but because the cost of errors there is the highest.

3. Five-Step Method for Designing Inspection Frequencies

Step 1: Grade the characteristics. Mark each control item in the CP as SC, CC, or general characteristic, and set the frequency according to the grade: use the highest frequency for SC characteristics and the lowest for general characteristics to avoid treating them equally. Grading should not be decided solely by the quality department; it should involve process and R&D to prevent downgrading SC characteristics to CC for convenience.

Step 2: Check the process baseline. If there are Cpk and control chart data, use the data to set the frequency: use the standard frequency for Cpk greater than 1.33, increase one level for 1 to 1.33, and rectify first for less than 1 before discussing frequency. For new processes without data, start with the strictest frequency and relax it after accumulating data. Use current process data, not the old Cpk from the PPAP submission, to set the frequency, or problems will eventually arise.

Step 3: Select the inspection method. The method determines the upper limit of the frequency: full inspection is suitable for high-risk items that can be automatically detected, and full inspection relying on human eyes should be used cautiously; sampling inspection is suitable for stable processes and destructive testing; first and last piece inspections cover the start of the shift, mold changes, and batch changes; patrol inspections combined with SPC control charts are suitable for continuous production processes, using trend monitoring to replace some sampling inspections. First and last piece inspections confirm "whether the status is correct," while sampling inspections monitor "whether there is any drift during the process." They complement each other and cannot be substituted for one another.

Step 4: Set the frequency and sample size. Provide a set of starting rules: the first piece must be inspected at the start of each shift, and the first piece must be inspected after mold or batch changes; SC characteristics should be inspected at least every 2 hours, and critical dimensions are recommended to use a c=0 plan; CC characteristics should be inspected 2 to 3 times per shift; general characteristics should be inspected once per shift. The sample size must be justified: use c=0 for critical characteristics and AQL for general characteristics. The patrol inspection interval should match the production rhythm, such as every 2 hours or every 200 pieces, and records must be filled out on-site, not after the shift. The frequency must be expressed in a clear and calculable manner; "sampling 5 pieces every 2 hours" is far more effective than "periodic sampling."

Step 5: Adjust as you go. Frequency is not a one-time decision: after running for a month, monitor three things—whether any nonconforming products have been missed, whether the inspection records match the actual on-site conditions, and whether the process capability has changed. If the process is under control for three consecutive months, the frequency for items other than SC can be relaxed; if the process fluctuates or Cpk declines, immediately increase the frequency. Each adjustment must be documented with the basis and date, providing a clear chain of evidence for why the frequency was changed from 4 hours to 2 hours. Include "when to relax and when to tighten" in the reaction plan to make the frequency dynamic.

Returning to the company at the beginning: the quality department changed the inspection frequency for SC characteristics to a c=0 sampling inspection every 2 hours plus first and last piece inspections, and reduced the frequency for general characteristics to once per shift. This change saved 20% of the inspection time and prevented any critical dimension nonconformities from flowing out within three months. Frequency design is not about being as strict as possible, but about allocating resources to the highest-risk areas.

4. Three Common Misconceptions

Misconception 1: Full inspection is the safest. Full inspection does not guarantee zero defects; human inspections can lead to fatigue and oversight, and only automated inspections can truly achieve full inspection. Using full inspection where poka-yoke and SPC should be used is a waste of money on minor issues. The correct sequence is to first ask, "can we use poka-yoke or automated inspection," and then decide whether to use full inspection.

Misconception 2: Frequency is fixed and unchanging. Not relaxing the frequency when process capability improves is wasteful, and not tightening it when the process fluctuates is a failure. Frequency must be adjusted based on data and changes.

Misconception 3: Write one thing and do another. If the CP specifies an inspection every 2 hours, but the records show it is done every 4 hours, the auditor will easily spot the discrepancy, and even the most reasonable frequency cannot save "inconsistency between what is said, written, and done."

5. In a Nutshell

Inspection frequency is the "breathing rhythm" of the control plan—the more critical the characteristic and the less stable the process, the more frequent the inspections should be. Setting the frequency based on data and adjusting it according to changes ensures both risk management and cost efficiency.

Inspection frequency is not a box to be filled in by guesswork, but a balance between risk and cost—set based on data and adjusted according to changes.

Knowledge code: 8.3.2

Version: v20260827

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