Inline vs Final Inspection: Two Gates for Better Handbag QC

Inline and final inspection are not alternatives; they answer different questions. Inline inspection happens while the run is still open, typically from day 12 of a 35-50 day bulk window with 20-30 units pulled from the line each day, and it catches drift while 60 percent of the order is still unmade. Final inspection happens when the lot is complete, drawing 125 units from a 1,200-piece lot at AQL 2.5 under ISO 2859-1, and it tells you whether to release the cartons. Inline inspection is the only QC activity that can still change the outcome of the run it is measuring; final inspection is the only one that describes the goods you are actually buying. A programme that runs both typically finds its defects at a cost of hours rather than at a cost of weeks.

Two gates, one comparison

A 3,000-Piece Tote Run That Passed Both Gates

An e-commerce brand in Manchester ordered 3,000 canvas totes in two colourways for a spring promotion, with a launch date that could not move. Sampling took 8 working days and the approved sample was signed in week three. Bulk was booked at 35-50 days with inspection pencilled for the last week. On day 11, an inline visit pulled 24 units from the line and found two things: the webbing handle was being set 4 mm short of the specified position, and the topstitch was running at 6 stitches per inch against a specified 7-9.

Neither finding was dramatic. Left alone, the handle position would have produced a bag whose drop measured 26 cm against a specified 30 cm, and the loose topstitch would have raised the seam-failure rate in the field. Because the run was only 11 days in, the correction cost one line reset and about 400 already-cut panels that had to be reworked. The final inspection at day 44 drew 200 units under ISO 2859-1 and passed with two minor defects against an accept number of 10. The launch held.

Three rules fall out of that run, and they generalise well beyond canvas totes. First, the value of an inline visit is proportional to how much of the run is still unmade, which is why day 11 beats day 30. Second, the defects worth catching inline are the ones that repeat, not the ones that are severe: a single bad unit is a sorting problem, whereas a drift is a setting problem. Third, the inline visit is only useful if it measures the same specification the final inspection will use, or the two gates will disagree.

What Inline Inspection Actually Catches

Inline inspection is process control, and its output is a correction rather than a verdict. The inspector walks the line, pulls units that have just been finished, and checks the operations that are expensive to repeat. Because the run is still open, a finding can be fixed in the remaining units, which is the entire economic case for the visit.

• Setting drift. Handle position, strap length and pocket placement measured in millimetres against the approved sample; a 3-5 mm drift is typical when a jig loosens over a long run.
• Stitch density and tension. Counted at 7-9 stitches per inch with Tex 40 thread; loose tension shows up as a visible loop on the underside and is the precursor to seam failure.
• Material substitution signals. A shade or hand-feel difference between panels cut in week one and week two, caught before the second dye lot is consumed.
• Hardware setting. Rivets and studs seated unevenly, or a plating batch that differs from the approved finish, which is far cheaper to stop than to re-sort.
• Branding registration. An embossed or printed mark drifting off its datum position by more than the specified 2 mm.

Inline visits are usually short and frequent rather than long and rare. Two visits of half a day each, at roughly 25 percent and 60 percent through the run, catch more than one full day at the end, because the first visit can still change the second half of the order. On a 500-piece order at MOQ level, even a single visit at the halfway point is worth more than a second final inspection, simply because there is still half a run to correct.

What Final Inspection Actually Catches

Final inspection is a verdict on a finished lot. It draws a statistically defined sample under ISO 2859-1, counts defects by class against the accept and reject numbers, and produces the document that decides whether the cartons move. It is the gate that a warehouse, a marketplace or a retailer can rely on, because it describes a defined population rather than a process.

• Workmanship across the whole run. Because the sample is drawn at random from the finished lot, it reflects every shift and every material lot that went into it.
• Function. Zippers cycled, closures worked, sliders and adjusters run through full travel, feet and studs checked for seating.
• Measurement. Body width, height, gusset depth and strap drop within plus or minus 3 mm on soft construction, 5 mm on structured bodies.
• Colour against the approved swatch. Delta E 2.0 or better under a D65 source, checked both within a bag and across bags.
• Packing and marking. Carton count, assortment, five-ply carton specification and shipping marks that match the documents.

What final inspection cannot do is improve the lot it measures. By the time the sample is drawn, every unit exists; the only remedies are sorting, rework, an allowance or a re-make, and the cheapest of those costs days. That asymmetry is the whole answer to the question of whether inline inspection is worth its fee.

The Comparison: Sampling Logic, Cost and Timing

Put side by side, the two gates differ in almost every dimension except the specification they both measure. The table below is the comparison worth pinning to the wall when a programme is being planned.

Read the last row carefully: the two gates are near-perfect complements. Inline is weak exactly where final is strong, and neither can substitute for the other. The reason to compare them is not to choose but to budget, because both are priced in man-days and the total QC spend on a first order should be a small, known line rather than an argument at the end.

When One Gate Is Enough

Sometimes it is, and pretending otherwise wastes money. A repeat order where the specification is unchanged, the material lot is the same and the previous two runs both passed is a reasonable candidate for final inspection alone. So is a very small top-up below the statistical scale, where a 100-percent check of 80 units costs less than designing a sampling plan.

1. Repeat with no change. Specification, material and tooling identical to a run that passed: final inspection alone is defensible.
2. Top-up under about 150 pcs. A full check may cost less than a sampling plan and produces a stronger result.
3. New silhouette or new material. Both gates, and an inline visit early rather than late.
4. New component source. Both gates, with the inline visit timed to the first day the new component reaches the line.
5. Season-critical launch. Both gates, with the inline visit written into the purchase order so it cannot be skipped when the line is busy.

The decision is really about how much of the run is still unmade. Where the answer is 60 percent or more, inline has something to protect. Where the answer is zero, only final remains, and the third-party inspection decision becomes the more useful lever.

How Two Gates Share One Specification

The failure mode of running both is disagreement: the inline visit tolerates a 4 mm variance and the final inspection rejects it, or the two use different reference samples. The fix is one signed specification that both gates measure against, with the tolerance stated in millimetres, the stitch density stated as a range, the colour threshold stated as a Delta E value with a named light source, and the defect classification agreed before either gate opens.

Reference artefact

One approved sample, signed and held, referenced by both gates

Dimensional tolerance

Plus or minus 3 mm soft, 5 mm structured, stated per dimension

Stitch density

7-9 stitches per inch, Tex 40 thread, 10 mm seam allowance, back-tacked

Colour threshold

Delta E 2.0 or better to the approved swatch under D65

Hardware

Zinc alloy, 24 h salt spray to ASTM B117, no base metal

Defect classes

Critical, major and minor defined with examples before inspection

With that document in place, an inline finding and a final finding are the same kind of statement, and the two reports can be read together as a trend rather than argued over separately. That is also what makes the AQL 2.5 numbers meaningful rather than abstract.

What to Ask For When Booking Both

Booking both gates is an administrative act, and the details decide whether the visits are useful. Ask for the inline dates to be fixed to production milestones rather than to calendar dates, because a visit booked for a Tuesday in week three is worthless if the line has not started cutting. Ask for the final inspection to be booked at about 80 percent packed where it is a pre-shipment inspection, so loose stock remains available for sorting.

• Fix inline to milestones. First visit at first finished units off the line; second at roughly 60 percent complete.
• Name the action limits. What triggers a line stop, and who has authority to call it.
• Require a written note within 24 hours of an inline visit, even where nothing was found.
• Book final at 80 percent packed, with the packing list supplied 48 hours ahead.
• Require both reports in the buyer's name, with photographs and a named inspector.

Programmes coordinated through our validated partner factory Quanzhou Yingming Bags run both gates against one specification, with export experience since 2014 and 7 production lines behind the scheduling flexibility that milestone-based booking requires. The full standard is set out in the handbag quality control reference.

The Failure That Only Two Gates Would Catch

The clearest argument for running both is the defect class that neither gate catches alone: a defect that appears midway through a run and disappears before the end. A component lot used for three days, a machine reset after a breakdown, or a substitute adhesive used while the specified one was out of stock all produce a small contaminated band inside an otherwise clean lot. A final inspection drawing 125 units from 3,000 may well miss a band of 150 units; an inline visit on the right day finds it immediately, because it is looking at what is being made right now.

That is the case in one sentence: final inspection is a photograph of the lot, and inline inspection is a video of the process. A photograph of a clean lot can still contain a band that never appears in the frame, and only the video shows it. Given that a single band of 150 defective units inside a 3,000-piece order can consume the margin on the whole shipment through returns, two short visits are among the cheapest insurance available to a private label handbag programme.

What Two Gates Cost Against the Order Value

Both gates are bought in man-days, so a small lot pays a higher percentage of its value than a large one even though the work is nearly identical. The figures below assume typical southern China day rates for one inline visit plus one final inspection. Read the last column against the alternative: a re-sort runs USD 0.35-0.60 per piece, so on a 500-piece lot the entire two-gate programme costs about the same as sorting half the lot once.

Placing the Two Gates on the Production Days

An inspection booked by calendar date rather than by production milestone arrives either too early to see finished units or too late to correct anything. The sequence below fixes each visit to a state of the run. Two external dates also belong on the same line: the container yard cut-off, usually two days before the vessel sails, and the booking itself, which has to precede the cargo-ready date by five to seven days in a busy month.

Action Limits for Inline Findings

Inline inspection only changes the outcome if someone can act without asking permission. Action limits convert a finding into a decision in advance, so the line does not wait for a coordinator in another time zone. Each limit below has three parts: what triggers it, what happens next, and who pays for the paused time. Writing the third part into the purchase order is what makes the first two real.

Frequently Asked Questions

People Also Ask