Featured Product
This Week in Quality Digest Live
Statistics Features
Douglas C. Fair
Part 3 of our series on SPC in a digital era
Scott A. Hindle
Part 2 of our series on SPC in a digital era
Donald J. Wheeler
Part 2: By trying to do better, we can make things worse
Douglas C. Fair
Introducing our series on SPC in a digital era
Donald J. Wheeler
Part 1: Process-hyphen-control illustrated

More Features

Statistics News
How to use Minitab statistical functions to improve business processes
New capability delivers deeper productivity insights to help manufacturers meet labor challenges
Day and a half workshop to learn, retain, and transfer GD&T knowledge across an organization
Elsmar Cove is a leading forum for quality and standards compliance
InfinityQS’ quality solutions have helped cold food and beverage manufacturers around the world optimize quality and safety
User friendly graphical user interface makes the R-based statistical engine easily accessible to anyone
Collect measurements, visual defect information, simple Go/No-Go situations from any online device
Good quality is adding an average of 11 percent to organizations’ revenue growth
Ability to subscribe with single-user minimum, floating license, and no long-term commitment

More News

John Niggl

Statistics

The Importer’s (Brief) Guide to Acceptance Sampling and Why It Matters

A basic understanding of acceptance sampling and AQL will help you get the most value out of inspection reports

Published: Tuesday, August 8, 2017 - 11:03

Ever wondered why quality control (QC) professionals check a sample instead of 100 percent of a shipment during inspection? Or maybe you’ve wondered why they use acceptance sampling, rather than simply inspecting an arbitrary quantity of goods, such as 10 or 20 percent?

Most importers value the transparency that quality control inspection provides. They know that catching any unacceptable quality issues or nonconformities before shipping is crucial to being able to address them before they hurt their bottom lines.

But many importers are at a loss when it comes to determining how many units they should inspect. They want to check as many as possible to get a representative look at their total order, while simultaneously trying to balance the time and cost needed for inspection. Frustration often results when importers don’t understand how standards dictate the sample size chosen for, and ultimately the results of, their inspection.

Luckily, if you count yourself among those importers that struggle to understand sampling, and acceptable quality limit (AQL) sampling in particular, it’s not as complicated as it may seem (see the ebook, The Importer’s Guide to Managing Product Quality With AQL). Let’s look at the significance of acceptance sampling, how AQL sampling factors into overall inspection results, and why it’s commonly used for QC inspection.

Acceptance sampling and where it came from

Acceptance sampling dates back to World War II, when the United States military needed a reliable and efficient way to test ammunition. The United States had been testing bullets for the war effort. Since it is necessary to destroy a bullet during the testing process and you obviously can't destroy every bullet, the military needed a way to test a sample of each batch, rather than 100 percent, so that most bullets in the shipment would be left intact for use in the field.

The solution was lot acceptance sampling, or simply, acceptance sampling. By using acceptance sampling to inspect a random sample of each batch of ammunition, the military was able to decide whether a batch of ammunition was likely to be acceptable. This concept soon led the U.S. Department of Defense to create Military Standard 105 (MIL-STD-105). Though largely discontinued in 1995, this standard formed the basis of most others which QC professionals continue to use to this day.

AQL sampling for quality control inspection

Quality control professionals have developed various sampling schemes or standards for product inspection over the years. Most sampling standards for inspecting consumer products use what are called acceptable quality levels—more recently called acceptable quality limits (AQL). AQL represents the minimum quality level (limit) that you, the customer, are willing to tolerate in an order of goods.

A number of sampling standards use AQL, including the more popular ANSI ASQ Z1.4, as well as ISO 2859-1 and other alternatives. Where inspection uses an AQL standard with random sampling, AQL largely determines if your shipment of goods passes or fails inspection.

Basics of AQL inspection

Most inspections of consumer products rely on a single-sampling plan. This means that inspection results are obtained by checking one random, fixed sample size of goods as determined by AQL.

When referring to a standard AQL table, you’ll find different AQLs correspond to different sample sizes set by your total order quantity and desired inspection scope, or inspection level.

Importers normally use three different AQLs for inspection, one for each of the three classes of defects generally defined as follows:

Critical—defects that violate regulations or pose a threat to user safety
Major—defects that would likely result in product returns but don’t pose a safety risk to the user
Minor—defects that are unacceptable in high quantities but generally won’t result in product returns

You’ll likely assign a lower, stricter AQL to more severe critical defects, while applying the highest of your three chosen AQLs to less serious, minor defects. But you may choose to classify particular defects differently than others in your same industry, depending on your product and your customers’ expectations.

A relatively strict AQL typically requires inspectors to check a larger sample of goods before arriving at a result that’s reliable and statistically significant. For example, an inspection of a lot size between 3,201 and 10,000 pieces with an AQL of 0.065 would require a minimum sample size of 200 pieces to return a clear result. Inspection of the same lot size with a more lenient AQL of 0.15 would only require a minimum sample size of 80 pieces.


Figure 1: AQL chart. Click for larger image

Each AQL and corresponding sample size is met with a limit to the number of defects an inspection of that sample can reveal without you rejecting the order (see “How Importers Use the AQL Table for Product Inspection”). Often called the rejection point, the AQL result will be a fail if you find a quantity of defects in excess of this number in the sample. You can use a free online AQL calculator to quickly find the AQL and corresponding rejection point and sample size.

How AQL sampling results affect inspection results

AQL sampling results are a major factor determining whether a shipment passes inspection, and consequently, whether you should accept the goods. But AQL sampling isn’t the only determinant. In most inspections using AQL, there’s an AQL result along with an overall result. And while a passing AQL result is a condition of an overall passing result, there are typically other conditions as well.

Despite finding defects in quantities and severities well within the tolerance you set with AQL, your order can still fail inspection for various reasons, such as:

Customer expectations are stricter than current AQL standards: This often happens when importers discover a particular quality issue is leading to a high number of product returns after they’ve already established inspection criteria. For example, an unexpected number of returns due to broken stitching in an item might prompt a garment importer to decide an order has failed inspection.

The product fails an important on-site test: Most products are subject to a regimen of on-site testing as part of inspection, depending on the product type. Some of these tests, such as hi-pot testing for electronics, are vital to ensuring the product functions safely and properly. Failure to pass such a test typically leads to a failing inspection, regardless of AQL results.

The product fails to meet industry regulations or retailer requirements: An order can fail inspection with a passing AQL result if the product doesn’t meet certain regulations or requirements. For example, the U.S. FDA has strict regulations for cosmetics products. Likewise, Amazon.com has requirements concerning poly bags and other packaging.

With all the rules importers must follow to use AQL sampling effectively, you may be wondering why you should bother with using AQL at all.

Most importers, especially those just beginning to manufacture, don’t clearly understand AQL sampling and the rules that govern it. But just because you may not fully understand AQL in practice doesn’t mean you can’t appreciate the advantages it presents over other sampling and inspection methods. There are a few reasons why AQL sampling remains the best approach for most importers.

Inspecting a sample is more efficient than inspecting 100 percent

The amount of insight that 100-percent inspection provides is unmatched. There’s no other way to see the quality and condition of more pieces in a shipment than by inspecting each and every one. That’s why some importers insist on 100-percent inspection, especially when placing their first order with a new supplier. They’ve decided that it’s too risky to inspect only a sample of their total order.

But inspection of every piece isn’t practical for the vast majority of importers. They have encroaching deadlines and commitments to meet in shipping their goods to their customers. So instead, they choose to inspect a random sample of the order because it’s more efficient. You can still make a reasonable estimation about the quality of your order relatively quickly by checking a representative sample.

Imagine trying to inspect 100 percent of 35,000 Bluetooth headsets. A thorough, professional inspection of this scale could take a team of three inspectors about 1.5 months to complete, assuming 250 pieces per inspector per day. How many importers have time to wait an additional 1.5 months before shipping to allow for QC inspection? More doubtful still is the prospect of your supplier willingly unpacking 35,000 finished headsets from packed cartons and repacking them after inspection.

Conversely, one to two inspectors using AQL sampling could report on the same shipment by checking a sample size of 125, 315, or 500 pieces in a single day. This approach would offer a smaller inspection scope and thus less confidence in the result than inspecting 100 percent. But you’d be able to make an informed shipping decision much faster. And if you’ve hired an independent third-party to inspect on your behalf, you’d likely spend much less money paying for their time.

Clear results backed by statistics

A common point of contention for importers concerning AQL sampling is the need to use specific sample sizes dictated by a sampling standard. Those unfamiliar with AQL often want to know why their inspector can’t simply check 10, 15, or 20 percent of the total quantity at random. Wouldn’t a round, whole percentage achieve similarly insightful results?

The short answer is no. Pulling an arbitrary, random sample does not achieve useful results in the same way as pulling a sample based on AQL. AQL sampling, along with other types of acceptance sampling, uses statistics to return a reliable result. In fact, obtaining a clear result is an important advantage of inspection that uses acceptance sampling. And in breaking from acceptance sampling standards, you lose this clear result that otherwise allows you to make an informed shipping decision.

Flexibility to change the standard at will

Another key advantage of using AQL sampling for inspection is that you can adjust quality tolerances anytime without needing to redefine what constitutes a “pass” or “fail” result. When inspecting 100 percent or an arbitrary percentage of an order, you’re forced to decide what types of defects and in what quantity of the inspected goods are acceptable. Any change to the sample size inspected forces you to rethink your tolerance.

In contrast, AQL sampling consistently offers a sample size and rejection point for any given AQL and order quantity. You’re free to tighten or loosen your tolerance for quality defects at any time. Many importers value this freedom because they often need to adjust quality expectations as they continue working with the same supplier.

You might begin with a relatively strict AQL of 2.5 for minor defects in the furniture you’re importing from Vietnam, for example. But if the first shipment fails inspection, you may find your quality standard is unreasonably high and choose to switch to a more lenient AQL of 4.0 for minor defects. Acceptance sampling affords you the flexibility to make this change virtually on the fly.

Conclusion

As an importer, you have a lot of options available for ensuring your products meet your requirements prior to shipping (see the e-book “How Experienced Importers Limit Product Defects in 3 Stages”). When it comes to preshipment inspection, a basic understanding of acceptance sampling and AQL will help you get the most value out of your inspection reports. And with greater insight into the quality of your products, you can take steps toward continuous improvement.

 

Discuss

About The Author

John Niggl’s picture

John Niggl

John Niggl is a client manager at InTouch Manufacturing Services, an American-owned company that provides solutions for quality and overseas manufacturing issues through product inspection and related quality control services.

Comments

AQL / RQL

Modern Acceptance Sampling plans use both AQL and RQL (with associated specified risks of incorrectly classifying the lot).  And accepting a lot in no way ensures that a lot is anywhere near as good as the AQL.  Acceptance only implies you are highly confident (depending on the consumer's risk chosen) that your quality level is better than RQL which of course is higher than AQL and much higher than the quality level needed.  Unfortunately, many users of acceptance sampling do not really understand the relatively low level of protection these plans actually provide (with typical sample sizes)

Bottom line, Acceptance Sampling should be viewed as an audit procedure to detect major issues that somehow went undetected.  If one wants to really control quality, it must be during the production process via SPC (rather than hoping to "find" defects in a sample after the product has been produced.

Just 1 more thing

Inspecting 100% of a lot is like adding a work station to the production line. Therefore the inpector becomes a production worker. And no matter how qualified the inspector is - will make mistakes. I've seen this in practice. Meanwhile taking a sample of a lot and usually taking it into a sterile environment gives the inspector the time and the peace of mind to actually "do the job" of inspection.