Knitted Bag Quality Control Plan: The Control Points, Hold Points and Records a Buyer Should Write into One

A Twelve Thousand Piece Order That Passed Inspection and Still Failed

A European retail buyer placed a twelve thousand piece knitted shopper in three colourways. The specification said AQL 2.5, the goods were inspected against AQL 2.5, the report came back with defects inside the acceptance number, and the shipment was released. Six weeks later the returns desk reported handles pulling away from the body on roughly one bag in forty.

The inspection had checked what an inspection checks: dimensions, visible defects, packing and count. It had not checked the thing that failed, because handle joinery strength is a laboratory measurement taken on a sample before bulk, not an attribute counted at goods-in. The order passed every check it was given and failed on the one check it was never given.

The post-mortem produced a plan rather than a complaint. Handle attachment became a control point with a named test method, a pre-bulk sample of five pieces per colourway, a minimum force figure and a written sign-off before knitting started. The same buyer now writes that clause into every knitted bag programme, and the failure has not recurred across four subsequent seasons.

What a Quality Control Plan Is Actually For

A quality control plan is the document that says which characteristic is checked, at which stage, by which method, against what limit, how often, and who signs the result. It is not a certificate, not an inspection report and not a specification. The specification says what the bag is; the plan says how each part of that description is verified and what happens when a check fails.

That distinction matters because most bag programmes have the first document and not the second. A buyer writes gauge, yarn count, dimensions, colour reference and hardware finish into a specification, then relies on a final inspection to catch anything that drifted. Final inspection catches the drift after it has been knitted into every piece in the lot, which is the most expensive possible moment to find it.

A plan moves the checks upstream. It puts yarn verification before knitting, gauge and grammage checks on the first-off panel, joinery strength before assembly, and dimensional checks before packing, so that a failure at any stage stops one stage rather than the whole lot. This is the entire commercial argument for writing one.

The Five Control Points on a Knitted Bag Line

A knitted bag programme has five places where a characteristic can be verified at low cost, and each one should appear in the plan with a method and a limit. Incoming yarn is the first: count, composition and colour against the approved reference, verified before a cone reaches the machine. Knitting is the second: gauge, stitch structure and course density on the first panels off the machine.

Assembly is the third, covering handle joinery, closures, linings and any bonded or sewn seam. Finishing is the fourth, covering steam shaping, blocking and any applied print or label. Packing is the fifth, covering count, folding, polybag, carton weight and marking. Every one of those five has a method that exists in the standards, and none of them is an opinion.

Control point 1 — incoming yarn

Count, composition, colour versus approved reference; verify before cones reach the machine

Control point 2 — knitting

Gauge, stitch structure, course and wale density, grammage on first-off panels

Control point 3 — assembly

Handle joinery strength, closure function, seam integrity, lining attachment

Control point 4 — finishing

Steam shape, blocking dimensions, print adhesion, label placement

Control point 5 — packing

Count per carton, folding method, polybag specification, carton weight and marking

Sample size at pre-bulk

Five pieces per colourway tested destructively, signed off before knitting starts

Sampling window

7 working days from approved brief to sample at the door

Bulk window

30 days from the event named in the order acknowledgement

Final inspection

AQL 2.5 under ISO 2859-1 with the general inspection level stated

Hold Points: Where Production Must Stop for Sign-Off

A control point becomes a hold point when production is not permitted to pass it without a signature. Three hold points are enough for most knitted bag programmes, and adding more slows the line without improving the outcome. The first is yarn release, the second is first-off panel approval, and the third is pre-shipment inspection release.

Yarn release is the highest value hold point and the one most often skipped, because yarn arrives in cones and looks correct. Verification is quick and inexpensive: count, composition and a shade check against the approved reference. A wrong yarn lot discovered after knitting has already consumed the yarn, the machine time and the finishing capacity, and none of the three comes back inside a 30 day window.

The Inspection Plan: AQL 2.5 and What It Does Not Cover

AQL 2.5 under ISO 2859-1 governs the final inspection, and it is the number most buyers already specify. It is a sampling scheme indexed by acceptance quality limit, and the sample size it produces depends on the lot size and the general inspection level, not on the AQL alone. A plan that states only 2.5 leaves the sample size to be decided at the dock.

The second thing a plan must say is what an inspection does not cover. Attribute inspection counts visible defects on the sampled pieces: dropped stitches, holes, soiling, broken hardware, misaligned print, wrong colour, packing errors. It does not measure abrasion resistance, tensile force, colour fastness or dimensional change, all of which are laboratory measurements taken on separate specimens before or during bulk.

So the plan needs two halves that are easy to confuse: the inspection plan, which is an attribute count at goods-in, and the test plan, which is a set of laboratory measurements on defined specimens. A buyer who writes both halves has covered the bag; a buyer who writes only the first has covered the visible defects and nothing that fails in week six.

Laboratory Tests That Sit Behind the Plan

Four laboratory measurements carry most of the risk on a knitted bag, and each has a real standard behind it. Abrasion resistance runs to ISO 12947 on a Martindale machine, with the end point agreed in cycles rather than left to the laboratory. Tensile force runs to ISO 13934-1 by the strip method or to ASTM D5034 by the grab method, and the two are not interchangeable.

Colour fastness splits into rubbing, laundering and light, covered respectively by ISO 105-X12 or AATCC 8, ISO 105-C06 or AATCC 61, and ISO 105-B02. Dimensional change runs to ISO 5077 using a washing procedure from ISO 6330. Each report should name the laboratory, the date, the method edition and the tested construction, because a result without a construction is not evidence about the bag being bought.

One knit-specific caution belongs in the plan. The Elmendorf tear method in ISO 13937 is written primarily for woven fabrics and is often unsuitable for knitted structures, where the tear path follows the loop rather than a straight line. Where tear resistance matters for a knit, a tongue or wing method, or a tensile and seam strength pair, usually produces the more defensible figure.

Records: What the File Must Contain at Goods-In

A plan is only as good as the file that proves it ran. At goods-in the buyer should be able to open one folder and find the approved sample record with its gauge and yarn count, the yarn release check, the first-off panel measurement, the pre-bulk laboratory reports, the in-line inspection records and the final AQL 2.5 report. Six documents, one folder, one order number.

The reason to insist on the full file rather than the final report is that the final report is the least diagnostic document in it. It says whether a lot passed; it does not say whether the yarn was the yarn, whether the gauge held, or whether the joinery was tested. When a retailer asks a question two seasons later, only the full file answers it.

A practical discipline: name the file structure in the order acknowledgement. Requesting six documents after delivery produces three; requesting them as a release condition produces six, because the goods do not move until the folder is complete. The lever exists at release and nowhere else.

Who Owns Each Line of the Plan

Every line of a plan needs an owner, and the owner is a role rather than a person, because people change and roles do not. Yarn release belongs to whoever issues the cones. First-off approval belongs to whoever sets the machine. Joinery testing belongs to whoever commissions the laboratory work. Final release belongs to whoever signs the inspection report, and for a branded programme that is often the buyer's own nominated inspector.

The buyer's own lines matter too. Someone on the buyer's side has to receive the pre-bulk reports and actually read them, and someone has to confirm that the general inspection level on the final report matches the one in the order acknowledgement. Those two checks take minutes and catch the majority of the mismatches that turn up in a quality dispute.

Reviewing the Plan After the Season

A plan written once and reused unchanged stops being a plan and becomes a formality. The review is short and specific: list every nonconformance raised during the season, note which control point should have caught it, and either move that check earlier or add the limit that was missing. Two or three changes per season is normal and is the sign that the plan is working.

Where a season produced no nonconformances at all, the opposite review applies. Either the plan is genuinely well placed, or the limits are loose enough that nothing can fail. Comparing the buyer's own returns rate against a zero-defect inspection record is the fastest way to tell which of those two is true.

Frequently Asked Questions

People Also Ask