Pro QC International provides global quality solutions. Click on your manufacturing country for more details.
Quality, supplier development and factory consulting solutions provided since 1984. We offer services in over 88 countries.
Pro QC has over 37 years experience in quality control and quality assurance across virtually every industry.
Questions about quality control inspection sampling are among the most frequently received.
With nearly four decades of inspections performed globally, we have taken a moment here to share a few of the most commonly noted questions and respective considerations.
• Which sampling plan should you use?
• How does it work?
• Do other sampling plan options exist?
• How do you determine the right sample size and AQLs?
Which sampling plan should you use?
The most commonly used QC inspection sampling plan is ANSI Z1.4-2003 Sampling Procedures and Tables for Inspection by Attributes, often referred to as an “AQL inspection.” AQL is the Acceptable Quality Limit that can be segmented by identifying major, minor, and critical non-conformances within a specified lot (batch). More specifically, the standard identifies AQL as “the quality level that is the worst tolerable process average.”
ANSI Z1.4 has a history with the U.S. Department of Defense and was originally known as MIL-STD 105E. The standard remains popular due to its ability to provide tightened, normal and reduced plans, including single, double and multiple sample options.
How does it work?
The basics include a lot (batch) that is represented as a quantity range column in the standard’s Table I. That number is matched with columns representing inspection levels. The result is a plotted code (represented as a letter) that identifies the sample size. Random samples totaling that number are evaluated throughout the lot, and a result is provided based on AQLs.
The lot (or batch)—
The lot is the total number of units. For us, it’s generally the shipment quantity of a particular SKU or family of SKUs. A consideration here would be in the case when there are multiple styles, colors, etc. The quantities can be combined to make the lot, or treated separately. For example, the sample size (and overall results) for one SKU of pink polka dot vests is going to be different than combining the pink polka dot and the animal print. That’s generally handled by dividing the total sample size out proportionately. If there are 20 pink polka dot and 80 animal print, 20% of the sample size is pulled from the pink polka dot style and 80% from the animal print.
The inspection levels—
The inspection levels “determine the relationship between the lot (batch) size and the sample size.” Straight from the standard:
“Three inspection levels: I, II and III are given in Table I for general use. Unless otherwise specified, Inspection Level II will be used. However, Inspection Level I may be specified when less discrimination is needed, or Level III may be specified for greater discrimination. Four additional special levels: S-1, S-2, S-3, and S-4, are given in the same table and may be used where relatively small sample sizes are necessary and large sampling risks can or must be tolerated. NOTE: In the designation of inspection levels S-1 to S-4, care must be exercised to avoid AQLs inconsistent with these inspection levels.”
An example of how this works includes:
Product ABC (Shipment Quantity 1000 Units)
Lot = 1000
Visual/Cosmetic/Workmanship – General Level II = “J” = 80 units
Functional (Destructive, Dimensional, Packaging Integrity, Etc) – Special Level 3 = “E” = 13 units
If you set the AQLs at 1.5 for major defects and 4.0 for minor defects, the inspector can go through findings from the random samples evaluated and determine a result. Using the visual/cosmetic/workmanship example noted above:
Major Defects @ 1.5 AQL:
Accept at 2 major defects found within the sample.
Reject at 3 major defects within the sample.
Minor Defects @ 4.0 AQL:
Accept at 5 minor defects within the sample.
Reject at 6 minor defects within the sample.
Critical defects can include safety issues, so the AQL is often 0.1, or no defects acceptable.
The above represents a single normal sampling example. Double and multiple sampling are an option here also, so one would see a smaller initial sample size and a second pull if within the AQL range. It’s possible to save time inspecting less, but it’s a substantial risk that the quality meets expectations. The double sample required would eliminate any time savings and potentially carry over to another day on-site.
According to Pro QC’s Operations Manager in China, Candace Wu, “the secondary sampling is rarely used for common daily products. According to statistics, the single sampling can meet the quality requirements of most products. Zero defects are usually required on the basis of very strict quality requirements for products by guests and end consumers.”
ANSI Z1.4 allows for reduced, tightened or normal inspections. As an example, when a supplier isn’t performing, they can be adjusted from normal to tightened for a period of time until corrective actions and subsequent improvements occur. These are considered under Switching Rules within the standard.
Note that a strong product specification and/or criteria is necessary for the result to be the most in line with expectations. Imagine major defects that should be classified as minor. With a reduced AQL for minor defects, the same shipment could pass or fail depending on the inspector’s interpretation. Photos of non-conformances should be included in inspection reports to avoid any subjectivity of the classification.
See the ANSI Z1.4 tables as a reference here: https://proqc.com/quality-resources/ansi-sampling-tables-web-application-standard/
Do other sampling plan options exist?
Another option includes a continuous sampling plan (MIL-STD-1235B) where you’re starting with 100% and after a certain amount is found OK, the sampling is reduced to one every specific number. If the inspector finds a defective unit, the sampling reverts back to 100%.
Also, the option of zero-based acceptance sampling exists as well. Within this sampling plan, no defects are allowed. An example would be zero-based @ 1%. In that lot example of 1000 units discussed before, the 1% would be a sample size of 13 units.
Pro QC works with clients of all sizes and within a number of industries. A large automotive manufacturer might reach out with specific zero-based sampling requirements, while a small business owner selling on Amazon might rely more on us to make a recommendation based on their stated needs.
How do you determine the right sample size and AQLs?
Determining the right sample sizes is a balance of things. Considerations include:
• The value of the shipment, usually taking into consideration the cost of the inspection. For example, a ANSI Z1.4 General Level II inspection of something might require two days whereas a reduced General Level I might be completed in one. This is a common question about how to balance cost and the number of samples evaluated. And, it should be noted here that reducing the sample sizes too much substantially increases the associated product quality risk. According to Pro QC’s Technical Supervisor for North America, Gerardo Trevino, “the sampling size can always be reduced in order to achieve the inspection in less time, but of course this has a substantial risk. Since the sampling was reduced, there are more possibilities that defects were not caught while performing the inspection.”
• Timing is always a consideration as well. Pre-shipment inspections are notorious for occurring the day or two before the shipment date. Suppliers don’t want to hold inventory for too long and arrange to be 80% produced and 100% packaged right before it’s time to head out. If that same example above has a two-day inspection requirement, and you have only one before it ships, that is a consideration. A third-party inspection company like Pro QC can often send multiple inspectors to complete the required sample in one day though. But, the consideration then becomes making sure the supplier has enough well-lit inspection area for multiple people to work effectively in.
• Sometimes, the destructive nature of the samples is a consideration. For example, if shell packaging must be damaged to check the function of a unit, the supplier must then re-pack or replace those samples.
No sampling plan reduces consumer risk to zero. However, conducting 100% inspections are most often too costly and time consuming. Thoughtful consideration into inspection levels and AQLs, as well as effectively classifying major, minor and critical defects is a sound way to balance to quality, risk and cost.
Contact Pro QC for additional information. Our teams of professionals are strategically located around the world and can assist with quality and engineering solutions on-site where and when you need support.
Supplier Audit - Advantages of Supplier Audits and Case Study
Supplier Audit - Advantages of Supplier Audits and Case Study...
Quality Management in the Indian Textile Industry
India's textiles industry is not only one of world's largest, spanning a diverse range of raw materials and textiles manufacturing outputs, but it is also, fascinatingly, on...
Pro QC's Global Team - A culture of excellence, exemplified by our people
A quality solutions provider is only as good as its team. Here at Pro QC, we are proud of our teams who work hard to solve the challenges faced by purchasing and quality d...
Quality assurance, supplier development and factory consulting services provided since 1984. We offer solutions in over 88 countries.
+1 813 252 4770
+52 81 2721 0928
+91 120 4291971
+33 9 7303 6784
+49 172 38 759 50
+886 2 2832 2990