This page is for anyone who has to answer a question that sounds odd the first time it is asked: how can checking only some of the pieces in an order tell you anything about the pieces nobody looked at. AQL, short for Acceptable Quality Level, is the answer the whole apparel industry has agreed on. It is a statistical inspection method: pull a sample of a known size from the whole lot, count the defects found in that sample, and compare the count against two fixed numbers, one that lets the lot pass and one that fails it.
This matters most to a quality inspector or a merchandiser at a factory, since final inspection before shipment almost never checks every piece in the order. Delta Garments, running Order A for Northline Apparel, is used throughout this course, starting with why its final inspection of this 15,000-piece T-shirt order pulls a sample rather than opening every carton.
Without this discipline, a factory either inspects every piece, costing time and money the order cannot absorb, or inspects too few pieces on no fixed rule, telling the buyer nothing reliable about the shipment it is about to accept.
You need: nothing but this page. You do not need to have run an inspection before.
By the end of this lesson you can:
Not covered here: how to read a sampling table to find the exact sample size and the acceptance and rejection numbers for a given lot (lesson 9.2.2), how defects are classified and counted (lesson 9.2.3), and what happens after a failed inspection (lesson 9.2.4).
| Term | Meaning |
|---|---|
| AQL (Acceptable Quality Level) | A statistical inspection method that checks a sample of goods and uses an agreed rule to pass or fail the whole batch. |
| Lot size | The total quantity of finished goods in the shipment or batch being inspected, before any sample is drawn from it. |
| 100% inspection | Checking every single piece in a lot, rather than a sample; the alternative AQL replaces for most final inspections. |
| Sampling table | The published reference table that turns a lot's own size into the exact sample size and pass/fail rule an inspector must use. |
| Final inspection | The last check of the finished, packed goods, done just before or at loading; a fail can stop shipment even at the last moment. |
A factory cannot open every carton of a 15,000-piece order and still ship on time. AQL solves this by fixing, in advance, exactly how large a sample must be for a lot of a given size, and exactly how many defects that sample may contain before the whole lot is rejected. The sample is not a handful of pieces an inspector happens to grab; it is a size taken from a published table, and the pass/fail rule is fixed before a single piece is opened, so the result cannot be adjusted after the count is known.
A person who has only seen one AQL inspection should not assume every one looks the same. What differs is who requires it and on what lot, not the underlying rule.
| What can differ | The shapes you meet | Why it matters |
|---|---|---|
| Export orders with a buyer-specified AQL | An export buyer such as Northline Apparel states its own required AQL level in the purchase order or the buyer manual, and often sends its own inspector or a third-party agency to draw the sample. | The factory cannot choose a looser AQL than the buyer specifies; the sample size and the pass/fail rule follow the buyer's own stated level. |
| Domestic orders inspected by the seller's own quality team | A domestic brand such as Cedar & Co., selling to its own shops from its own forecast, still runs an AQL-style final inspection on its own finished stock before it leaves the factory, even with no outside buyer to satisfy. | The same sampling logic protects the seller's own reputation with its own customers, not only a contractual buyer relationship. |
| A lot made to order (MTO) versus a lot made to stock (MTS) | An MTO export lot, like Order A, is inspected once, against the one buyer's specification, before the one confirmed shipment leaves. An MTS domestic lot is inspected in batches as stock is finished, against the seller's own standard, before each batch reaches the shops. | The AQL method is identical either way; only who sets the standard and when the inspection happens differs. |
| Product families beyond a basic garment | The same sampling logic applies to socks and hosiery counted by the dozen pair, to towels and home textiles counted by the piece, and to knit-to-garment styles, not only to a cut-and-sewn T-shirt or a pair of jeans. | A learner who has only inspected garments should expect the same lot-to-sample logic on any countable finished product, with the lot size simply defined in that product's own unit. |
Order A is Delta Garments' 15,000-piece, single-colour T-shirt order for Northline Apparel, made to order for one confirmed export shipment. Before the goods are packed and loaded, Delta Garments' quality team runs a final inspection under AQL, agreed with Northline Apparel in the purchase order. Northline Apparel does not ask Delta Garments to open all 15,000 pieces; it asks for a sample drawn under the sampling table's own rule, because a correctly drawn sample of this size gives Northline Apparel a statistically reliable picture of the whole shipment.
| Fact | What it tells the inspector |
|---|---|
| The lot size is Order A's own 15,000 finished, packed pieces | The lot size, not a guess, is what the sampling table is entered with |
| Northline Apparel's purchase order states the required AQL level | The factory does not choose its own AQL level; the buyer's contract fixes it |
| The sample must be drawn at random from across the whole lot, not from one carton | A sample taken from only the first cartons packed does not represent the whole 15,000 pieces |
| The pass/fail rule is fixed before the sample is opened | Nobody adjusts the rule after seeing how many defects turn up |
Order A being made to one buyer's confirmed specification, for one shipment, is exactly why its AQL level and its sample come from Northline Apparel's own stated requirement rather than from Delta Garments' own general practice.
The worked numberThis example uses Order A's own fixed lot size of 15,000 pieces; the exact sample size Delta Garments must pull is worked out from the sampling table in the next lesson. What this lesson shows is the shape of the saving: a sample is a small share of the lot, not a small handful chosen without a rule, and it still lets the inspector judge the whole 15,000 pieces.
A common mistake is an inspector pulling a dozen pieces from the top of the nearest cartons, calling this "a quick AQL check," and reporting the lot as passed. A sample this size, drawn this way, is neither the size the sampling table requires for a 15,000-piece lot nor drawn randomly across it, so its result says nothing statistically reliable about the other pieces. A lot passed on this basis can still ship with a real defect rate the sample never had the size or the spread to detect.
AQL is a way of checking a big shipment of goods without opening every single piece. An inspector takes a sample of a size fixed by a published table, counts the faults found in it, and compares that count against two numbers fixed in advance: one that lets the whole shipment pass, one that fails it. A properly sized, randomly drawn sample can stand for a much larger shipment. A small handful of pieces grabbed without a rule cannot, even if someone calls it a quick check.
Mona, the merchandiser at Delta Garments, is preparing Order A for final inspection before shipment to Northline Apparel. The floor supervisor, keen to finish quickly before the port cut-off, suggests opening only the two cartons nearest the packing table and counting whatever faults turn up there.
The supervisor's note reads:
Should Mona let the floor supervisor treat two conveniently placed cartons as the AQL sample for Order A's final inspection.
No. Two cartons chosen for convenience are neither the size the sampling table requires for a 15,000-piece lot nor a random sample across it, so a result from them is not a valid AQL result.
First, the sample size is fixed by the lot size, not by how much time is left. Order A's own 15,000-piece lot requires a sample size read from the sampling table, worked out in the next lesson; two cartons is very unlikely to match it.
Second, the two cartons were packed from the earliest sewn pieces, not drawn at random. An early run can differ from later production if a machine setting, a fabric roll or an operator's pace changed partway through the order, so an early-only sample can miss a fault that only appears later in the run.
Mona should have the sample drawn at random across the full lot, sized from the sampling table, even under time pressure. Reporting a convenience sample as an AQL pass risks shipping a lot Northline Apparel never actually agreed to accept on that basis, and a later claim of a false pass is far costlier than the time saved.
If Delta Garments ships Order A on the strength of a two-carton convenience sample and Northline Apparel later runs its own proper AQL check on arrival and finds a real defect rate the two cartons never revealed, Northline Apparel can reject the shipment on arrival rather than at the factory gate. Rejecting 15,000 pieces after they have already been shipped costs the freight already paid, plus the time to rework or replace the goods, plus the risk to the next order from the same buyer.
Compared with the small amount of extra time a correctly sized, correctly drawn sample takes at the factory, the cost of a false pass discovered after shipment is far larger, because by then the goods, the freight, and the time are already spent.
Open the AQL final inspection report builder. For now, look only at the Lot size and Sample drawn columns: Order A's lot size is already entered, but the sample size itself is left blank, because it is calculated from the sampling table, which the next lesson covers in full. This lesson's job is only to show why those two columns exist and why the sample column cannot be filled from a guess.
1. What does AQL let an inspector do?
2. Why does a two-carton sample taken only from pieces packed early in Order A's run fail to be a valid AQL sample?
3. A domestic brand producing to its own stock, with no outside buyer to satisfy, still runs an AQL-style final inspection before goods reach its shops. Why?
| EN | AR | FR | TR | Meaning |
|---|---|---|---|---|
| AQL (Acceptable Quality Level) | مستوى الجودة المقبول (AQL) | Niveau de qualité acceptable (NQA) | Kabul edilebilir kalite düzeyi (AQL) | A statistical inspection method that checks a sample of goods and uses an agreed rule to pass or fail the whole batch. |
| Lot size | حجم الدفعة | Taille du lot | Parti büyüklüğü | The total quantity of finished goods in the shipment or batch being inspected, before any sample is drawn from it. |
| 100% inspection | الفحص الكامل (100%) | Inspection à 100 % | %100 muayene | Checking every single piece in a lot, rather than a sample; the alternative AQL replaces for most final inspections. |
| Sampling table | جدول أخذ العينات | Table d'échantillonnage | Örnekleme tablosu | The published reference table that turns a lot's own size into the exact sample size and pass/fail rule an inspector must use. |
| Final inspection | الفحص النهائي | Inspection finale | Nihai kontrol | The last check of the finished, packed goods, done just before or at loading; a fail can stop shipment even at the last moment. |
You now know why a factory samples instead of inspecting every piece. The next lesson opens the sampling table itself: how a lot's own size tells the inspector exactly how many pieces to pull and how many defects the sample may still pass with. Back to the course map