Jacket Quality Control: From Inline Checks to Final AQL Inspection

What the Full Jacket QC Process Looks Like

The direct answer is short: a jacket order should pass through four control gates, and each gate must produce a record you can audit. The gates differ in purpose. The first protects you from bad inputs. The second protects you from repeating a defect across hundreds of units. The third catches what machines and operators missed before the garment is packed. The fourth proves, with a defined sampling method, that the lot as a whole meets the specification you signed.

• Gate 1 - Pre-production: fabric, lining, down or insulation, zippers, threads, labels and trims are checked against approved standards before cutting begins. Shade bands, shrinkage, weight per square meter and hand feel are verified against the sealed standard.
• Gate 2 - Inline: the first piece off the line is approved against the tech pack, then inspectors rove the line at a defined frequency to catch seam, tension, alignment and assembly errors while the batch is still small.
• Gate 3 - End-of-line: every garment is checked for workmanship, measurements, hardware function, cleanliness and labeling before it reaches the packing table.
• Gate 4 - Final AQL: a random sample is drawn from the finished lot and inspected against a defined defect classification. The lot is accepted or rejected according to the acceptance and rejection numbers for the chosen AQL.

Each gate should output something written: an incoming material report, a first-piece approval sheet with a signature and date, an inline defect log, a measurement record, and a final inspection report with photographs. If a factory cannot show these records, you are relying on memory and goodwill rather than a system. The records also matter commercially, because they are the evidence you use when a claim arises months later.

One structural point is often missed. Gates 1 to 3 belong to the factory and should be running whether or not you are watching. Gate 4 is the buyer-facing proof, and it is the one most brands pay a third party to witness. A factory that only performs well when an auditor is present has a supervision problem, not an inspection problem. When you evaluate a new supplier, ask to see the inline logs from a recent order rather than the polished final report. Browse our jacket product range to see the categories these controls apply to, from down and puffer styles to softshell and varsity programs.

Why AQL 2.5 Is the Standard for Jacket Final Inspection

AQL stands for acceptance quality limit. It is the worst tolerable process average, expressed as a percentage of defective units, that a sampling plan will still accept most of the time. AQL 2.5 does not mean a factory is allowed to ship 2.5 percent defective jackets. It means that under the sampling plan, a lot with roughly 2.5 percent major defects has a defined probability of acceptance, and a lot meaningfully worse than that will usually be rejected. The distinction matters when a supplier says "2.5 percent is within AQL" as an excuse. It is a sampling boundary, not a quality target.

The system comes from ISO 2859-1, which grew out of the old ANSI/ASQ Z1.4 tables. It defines sample sizes by lot size and inspection level, and acceptance numbers by AQL. For apparel, the common configuration is general inspection level II, normal severity, single sampling, with major defects at AQL 2.5 and minor defects at AQL 4.0. Critical defects, meaning safety or regulatory failures, are normally set at AQL 0, so a single occurrence rejects the lot regardless of the sample size.

Three practical consequences follow. First, your purchase order must name the standard, the inspection level, the AQL values and the defect definitions. "Final inspection required" is not a specification. Second, the sample must be drawn randomly from the packed or ready-to-pack lot, not selected by the factory. Third, the result is only meaningful if the lot is homogeneous: one style, one colorway, one production window. Mixed lots distort the sample. For a deeper look at how these layers connect inside a factory, read how quality control works in jacket factories.

QC Gates Mapped to Your Production Calendar

Quality control is a scheduling problem as much as an inspection problem. Each gate needs a slot, an owner and a decision deadline. If the inline gate slips, the final inspection inherits defects that can only be fixed by rework, and rework is where delivery dates are lost.

For a program with 7 to 14 days of sampling and 25 to 40 days of bulk production, the pre-production gate should close before the cutting ticket is released, inline checks should run daily once sewing starts, and the final inspection should be booked two to three days before the ship date so there is room to react. Brands that book the final inspection on the loading day are buying information, not protection.

It also helps to separate the two kinds of inspection by purpose. Inline inspection is diagnostic: it exists to find and fix a cause. Final inspection is judicial: it exists to accept or reject a lot against a standard. Confusing the two leads to factories arguing that a defect "was already there" and buyers arguing that the final report should have caught it earlier. Write both into the quality agreement, and make the final report reference the inline log so the history is visible. Our guide to private label manufacturing quality standards covers how to phrase those obligations.

How AQL 2.5 Sampling Works on a Real Jacket Lot

The sample size is not a matter of opinion. It comes from the lot size and the inspection level. Under general inspection level II and normal severity, a lot of 26 to 50 pieces draws 8 units, and the acceptance number for AQL 2.5 is zero: no major defects are allowed in those 8 units. That single fact changes how small orders should be planned. At the 50-piece minimum order quantity we run, the final inspection is a zero-defect gate on a small sample, so the inline work has to be honest.

Read the table as a planning tool, not a menu. At 500 units, 50 jackets are inspected and one major defect is still acceptable. At 5,000 units, 200 are inspected and five major defects are acceptable. The absolute number of tolerated defects grows with the lot, which is why buyers who need a consistent defect rate across sizes should split lots or tighten the AQL rather than rely on one large sample. For the economics of order size, see how MOQ affects jacket unit cost.

Two mechanics are worth knowing before you argue with an inspector. First, switching rules: normal severity is the default, but ISO 2859-1 allows tightened inspection after repeated rejections and reduced inspection after a run of acceptances. If your supplier's reports never change severity, the plan may be decorative. Second, sampling is not a substitute for 100 percent checks on safety features. Drawcords, small parts and labeling compliance should be verified on every unit or at least on every carton, because a single escape can trigger a recall that no AQL number will cover.

[IMAGE_PLACEHOLDER]

Photo prompt: a jacket final inspection station in a Chinese outerwear factory, inspector measuring the chest width of a navy puffer jacket on a lit table, measurement tape and AQL inspection report visible, other jackets stacked behind, natural daylight, documentary style, no logos.

Critical, Major, and Minor: The Classification That Decides Pass or Fail

AQL only works when everyone agrees what counts as a defect. The classification does the work. A critical defect is a safety, regulatory or legal failure that makes the product unsafe or unsellable, and it is normally judged at zero acceptance. A major defect is a functional or visible failure that a customer would reject or that would make the garment unsellable at full price. A minor defect is a cosmetic deviation that does not affect wear, function or the stated specification. Ambiguity here is the most common reason a final inspection becomes a negotiation.

Two boundary rules prevent most disputes. First, define measurement defects by the tolerance, not by opinion: if the chest measures outside the allowed range, it is a major defect even if the garment looks fine on the table. Second, define repeatability: three minor defects on the same garment or in the same critical area can be counted as one major, because the cumulative impression is a visible quality failure. Both rules should be explicit, and both should be illustrated with photos in the annex so a new inspector applies them the same way.

What Inline Checks Catch That Final Inspection Cannot Fix

Inline inspection is the only gate that can prevent a defect from being repeated. By the time a jacket reaches final inspection, the error has already been replicated across the batch, and the only remedies are rework, replacement or concession. That is why the inline gate deserves more attention than it usually gets in buyer-side quality plans.

• First-piece approval: the first garment from each operation is measured and compared with the tech pack before the line runs at speed. Approval is signed by the line supervisor and the QA lead, and the approved piece stays at the station as a visual reference.
• Roving checks: an inspector visits each workstation at a defined interval, checks the last units produced, and records the result. The interval should be short enough that a defect is caught within tens of units, not hundreds.
• Needle and sharp-tool control: broken needles are logged, the garment is quarantined and checked, and needle fragments are accounted for. This is a safety control, not a housekeeping detail.
• Seam and tension checks: stitch density, seam allowance, puckering and skipped stitches are checked on curved seams, armholes and hood attachments, where most workmanship failures begin.
• Down and insulation containment: quilting channels, baffle seams and needle holes are checked for leakage after filling, since leakage is difficult and expensive to correct later.
• Hardware function: zippers, snaps, cord locks and adjusters are cycled at the station, not only at the end of the line, so a bad batch of hardware is identified before it is installed across the order.

A practical inline record has four fields: time, station, finding, and action. If the log shows only a row of ticks, it is not evidence of inspection. If it shows findings and corrections, it is. Ask for the log from a completed order and look at whether the same defect appears repeatedly at the same station; that pattern points to a training or equipment problem that a final inspection would only punish, never solve.

The same logic applies to subcontracting. If parts of the jacket, such as embroidery, quilting or seam sealing, are outsourced, the inline gate should follow the work. A factory that inspects only what happens inside its own walls is not controlling the product. Our private label jacket manufacturing process explains how the stages connect from pattern to packing.

Measurement Checks and Points of Measure

Fit failures are the most expensive defects to discover late, because they affect every size in the run. Measurement inspection should be a fixed list of points of measure, measured on a flat table after the garment has rested, with the tolerance written next to each point.

• Chest or bust, one inch below the armhole: the primary fit dimension and usually the tightest commercial tolerance.
• Body length, from high point of shoulder to hem: check against the graded spec for each size, not only the base size.
• Sleeve length and cuff opening: critical for outerwear, where a short sleeve exposes the wrist and a tight cuff blocks layering.
• Shoulder width and armhole depth: the two dimensions most often distorted by inconsistent seam allowance.
• Hood height and opening: affects both fit and safety compliance on childrenswear.
• Pocket placement and opening width: a visible mismatch between left and right is a major defect even when each pocket is individually within tolerance.
• Hem width and side seam length: catches twisting and cut-and-sew alignment problems.

As a starting point, many tech packs use plus or minus 1.0 cm on major dimensions such as chest, body length and sleeve, and plus or minus 0.5 cm on smaller dimensions such as collar, cuffs and pocket placement. These are commercial conventions, not universal law. Confirm the values for your product, and state whether the tolerance applies before or after washing, because a shell fabric with high shrinkage can pass on the table and fail after the first home wash. If your program includes performance fabrics, the interaction between membrane, lining and seam construction also changes how the garment measures and hangs; see our notes on ski jacket manufacturing and performance specs.

Measurement records should list the actual value, not only pass or fail. Actual values let you see drift: if the chest of every unit is 0.8 cm under nominal, the pattern or the cutting allowance is wrong, and the next order will repeat it. A report that shows only ticks hides the trend that matters most.

Functional Tests Buyers Should Require

Visual inspection alone will not tell you whether a jacket works. A short functional protocol, applied to the sample and repeated on the final lot, catches the failures that generate returns.

• Zipper cycling: open and close the main zipper 20 to 30 times, including around the curve at the hem and collar. Record jamming, separation, skipped teeth and slider drag.
• Seam strength pull: apply a firm manual pull to side seams, armhole seams and pocket attachments. Look for stitch rupture and seam grin, especially on lightweight shell fabrics.
• Down leakage check: shake and compress the garment, then inspect seams, quilting lines and needle holes against a dark surface. A few fibers are normal; visible clusters of down indicate a containment problem.
• Water repellency: apply a light spray to the shell and observe beading and wet-out. This is a screening test, not a lab replacement, but it catches missing or inconsistent DWR treatment.
• Fill weight verification: weigh a sample garment and compare against the specified fill weight, allowing for the shell and trim. Underfilling is a common silent substitution in down programs.
• Hardware and cord safety: cycle snaps, cord locks and adjusters, and check cord end lengths and attachment strength against the applicable safety standard for the target market.

Write the tests, the number of cycles and the acceptance criteria into the inspection annex. "Zipper works" is not a criterion; "main zipper cycles 25 times without separation, jamming or tooth damage" is. The difference is enforceable.

[IMAGE_PLACEHOLDER]

Photo prompt: close-up of a quality inspector testing a black jacket zipper by cycling the slider, gloved hands, measuring tape and defect log on the table, factory sewing floor blurred in the background, neutral lighting, editorial photography, no branding.

How to Read a QC Report and What to Challenge

A final inspection report is a small dataset, and most of its value is in the parts that are easy to skip. Read it in a fixed order so a polished summary does not hide a weak sample.

• Lot definition: style, color, size range and quantity inspected. If the report covers a mixed lot, the sampling math is invalid.
• Sample size and level: confirm the sample matches the lot size under the stated inspection level, and that the sample was drawn randomly.
• Defect counts by class: compare against the acceptance and rejection numbers. A report that lists defects but not the AQL decision is incomplete.
• Measurement table: look for actual values and the number of units outside tolerance. One out-of-tolerance unit in a sample of eight is already a rejection at AQL 2.5 for a 50-piece lot.
• Photographs: defects should be shown with a scale reference and a legible tag. Photos of the packing area are not evidence of workmanship.
• Carton and packing check: carton drop, label placement, polybag warnings and size ratio accuracy. Retail chargebacks often come from packing errors, not sewing errors.

Challenge three things routinely. First, a report issued by the production manager rather than an independent QA function. Second, a report with no measurement values. Third, a report dated after the container was loaded. Any of the three means the inspection was documentation, not control. For the commercial side of inspection timing, including who pays for re-inspection, see avoiding hidden costs in jacket sourcing.

When a Lot Fails: Sorting, Rework, and Re-Inspection

A failed final inspection is not the end of a program, but it is a schedule event. The lot cannot ship until the defect is corrected and the correction is verified. The sequence below is the one that keeps the delay measurable instead of open-ended.

Typical impact: 3 to 10 days for a sort and re-inspection on a mid-size jacket lot, depending on whether repairs are needed and whether the defect is confined to one operation.

Agree the cost rules before the order starts. Who pays for the sort, the re-inspection and the freight delay? The usual answer is that the party responsible for the defect bears the direct cost, but that only works if responsibility is defined. A clause that says "defective goods will be replaced" is weaker than a clause that says "if the final inspection is rejected, the supplier will sort 100 percent at its own cost within five working days, and a re-inspection will be performed at the supplier's cost before shipment." The second version can be scheduled.

There is also a commercial option that buyers forget: partial acceptance with a price adjustment. For a minor-only failure where the garments are sellable at a discount, a concession can protect the season. For any critical defect, concession is not appropriate, because the risk sits with the final customer and the brand, not with the factory.

What QC Adds to Timeline and Budget

Quality control has a visible cost and an invisible benefit. The visible cost is inspection time, sample units destroyed or retained, and sometimes a third-party inspector. The invisible benefit is the avoided cost of returns, markdowns, chargebacks and emergency air freight.

On the calendar, plan for the gates rather than around them. Sampling typically takes 7 to 14 days, and bulk production 25 to 40 days for jacket programs. Inline checks run inside the bulk window and do not add days if they are scheduled with production. Final inspection needs one to two days plus a buffer for corrections. A realistic plan books the final inspection two to three days before the ex-factory date, not on the same day. If a third-party inspector must travel, add scheduling lead time, especially in peak season when inspection companies are heavily booked.

On the budget, the main controllable items are sample cost and rework risk. Sample fees for jacket development typically fall in the range of 50 to 200 US dollars depending on complexity, and are commonly credited against the bulk order. That fee buys the fit and construction decisions that determine whether the bulk lot passes. Skipping or rushing sample approval is the most expensive saving in the whole program, because every measurement and workmanship issue that survives the sample is replicated across the order.

Payment structure interacts with quality leverage. A 30 percent deposit and 70 percent balance before shipment is standard for private label outerwear. The balance is your last point of leverage, so tie it to the final inspection result rather than to the loading date. If the inspection is failed, the balance should not be released until the re-inspection passes. That single sentence in the payment terms does more for quality than most audit programs.

Common QC Failure Modes in Jacket Programs

Most failures repeat across the industry, and most are preventable at the specification stage. The table below maps the failure to its usual cause and the control that addresses it.

Notice how few of these are sewing-skill problems. Most are specification, material or communication problems that surface as workmanship defects. That is why a quality conversation that starts at the sewing line is already late. It should start with the tech pack, the approved sample and the material standard. Our article on how down jacket manufacturers ensure consistency goes deeper into the process controls behind this list.

Putting QC Requirements Into the Purchase Order

The purchase order is the only quality document that both sides have already agreed to. Attach a one-page inspection annex and reference it in the PO body. The annex does not need to be long, but it needs to be specific.

1. Standard and plan: state ISO 2859-1 or ANSI/ASQ Z1.4, general inspection level II, normal severity, single sampling.
2. AQL values: majors at 2.5, minors at 4.0, critical at 0, unless your category requires tighter limits.
3. Defect definitions: list examples by class, with photos, and state that safety-related defects are always critical.
4. Tolerances: provide the points of measure and the allowed range for each, and state whether they apply before or after washing.
5. Sample and golden standard: name the approved sample, keep a sealed copy with the factory and a duplicate with you, and state that bulk must match it.
6. Access and timing: the buyer or its nominated inspector may inspect at any reasonable time, and final inspection happens before packing is closed.
7. Failure handling: define sort, rework, re-inspection, cost responsibility and the maximum allowed delay.
8. Payment linkage: balance released after the final inspection passes, not after loading.
9. Records: the factory provides the inspection report, measurement values and defect photos within two working days.

Keep the annex stable across orders and version it when the product changes. A stable annex becomes a shared language: the factory knows what will be measured, the inspector knows what to count, and you know what you are paying for. When a dispute arises, the conversation moves from opinion to the document, which is exactly where it should be. If you need help drafting the annex for a specific jacket category, send us your requirements and we will review the tech pack against the inspection plan.

How Ginwen Runs QC for Private Label Jackets

Ginwen is a jacket and outerwear factory in Dongguan, China, producing down jackets, puffer jackets, cotton-filled jackets and technical outerwear for brands on OEM, ODM and private label programs. Our quality system follows the gate structure described above: incoming material checks, first-piece approval, inline roving inspection, end-of-line checking and a final AQL inspection before packing.

The operating facts matter more than the adjectives. Our minimum order quantity is 50 pieces per style. Sampling takes 7 to 14 days, and bulk production 25 to 40 days depending on complexity and season. Sample fees range from 50 to 200 US dollars and are credited against the bulk order. Payment terms are 30 percent deposit with 70 percent balance before shipment. We hold ISO 9001, BSCI, RDS and OEKO-TEX certifications, and we can provide current copies scoped to the producing facility during supplier evaluation.

What that means for a buyer is practical. A 50-piece order is small enough to test a new style and large enough to run a real AQL gate, so the first order doubles as a quality trial. The inline records from that order tell you how the factory behaves before you commit to a seasonal program. If the style grows, the same inspection annex scales to larger lots without rewriting the process.

We do not publish invented client names, certification numbers or market statistics. If you need a reference for a specific category or market, ask, and we will tell you what we can share and what we cannot. That honesty is part of the quality system too: a supplier that overstates its capability at the quotation stage will usually overstate its inspection results later. You can review our product categories or contact the team to discuss a program.

[IMAGE_PLACEHOLDER]

Photo prompt: wide shot of a clean jacket factory inspection area in Dongguan, several padded jackets hanging on a rail, two inspectors at a long white table with measurement sheets and a laptop, cartons labeled for export in the background, bright industrial lighting, realistic documentary style, no visible brand logos.

Frequently Asked Questions