Home › Field notes › Sample Revision Rounds Management: Criteria, Limits and Delivery Impac

Modular Bag Pro · 2026-10-11 · 2,491 words · 11 min read

A sample revision round is one complete rebuild-and-review cycle: a mark-up is issued against acceptance criteria written beforehand, the piece is rebuilt to that mark-up, and a decision is recorded. The first round sits inside the sampling fee of USD 50-150, and a second repeats the setup, the short-length material purchase and the courier dispatch of the first. Prototypes are built in 6-10 working days and in 12-15 where the construction is complex, the order floor is 500 units per reference, and bulk then occupies 35-50 days, so every further round consumes calendar that was reserved for production. Figures are indicative and quoted FOB Xiamen for civilian load-carriage goods; this is general trade information.

What Counts as One Revision Round

The word round is used loosely in most programmes, and the looseness is what later turns into an argument about fees and dates. A round has three elements and it is only a round when all three are present: a mark-up issued in writing against the specification, a physical rebuild to that mark-up, and a recorded decision afterwards. Anything missing one of those three is correspondence, not a round.

An email asking whether a pocket depth can be 20 mm instead of the drawn value is not a round. Nor is a request for photographs, a measurement clarification, or a question about whether a component is available in another finish. Those are queries, they cost answering time, and they should be answered inside the current round rather than starting a new one.

The written acceptance criteria are the part most programmes skip, and they are the part that makes a round finite. If the specification says only as discussed, then every review is an opinion and the review will simply continue until somebody gets tired. Criteria written before the build — dimensions with tolerances, function checks, attachment fit, fastener operation — turn the review into a pass or fail exercise with a list.

Every round must end with one of three recorded outcomes: approved, approved with deviations noted in writing, or refused with a fresh mark-up attached. A round that ends with a phone call and no written outcome has not ended, and it will be reopened later by whoever was not on the call.

Takeaway: Treat a round as beginning only when a written mark-up is issued and ending only when a written decision is recorded, because a round without those two documents keeps running and quietly consumes the production calendar.

Round-by-Round Acceptance Criteria

Criteria should tighten as rounds progress, for a simple reason: the first round exists to find out whether the specification was right, and later rounds exist to find out whether the corrections were made. Those are different questions and they deserve different tests. Applying the same loose checklist three times produces three rounds that all pass and a product nobody is happy with.

One rule keeps this honest: a criterion that was not written before the build cannot be introduced at the review. New requirements raised at the review are a specification change, and they belong in a fresh mark-up rather than being folded into the round under discussion.

Verdict: Write the acceptance criteria before each build and refuse any criterion introduced for the first time at the review, because a criterion invented during review converts a finite round into an open-ended one.

Why One More Change Costs More Than the Change

The cost of a revision is not the cost of the alteration; it is the cost of the rebuild around it. Moving a pocket by a modest distance means a new marker, fresh cutting, a re-sewn shell, refitting of any hardware in the affected area, and a fresh dispatch. The alteration itself might occupy minutes of sewing time, but everything around it repeats in full.

The queue is the second cost and the one buyers never see. A build slot was reserved for the original piece; when a second round is needed, that slot is gone and the new build takes the next available one. On a floor with 7 production lines running 149 machines, the next slot is a function of everything else booked at that moment, not of how small the change is.

Tooling can be drawn in, which changes the order of magnitude. A revision that alters the curvature of a part where a decoration sits re-opens the pad or die; a revision that alters a moulded component re-opens the cavity. Those carry one-off charges in the indicative USD 300-2,500 band, and the revision stops being a sampling matter altogether.

Decision quality is the quietest cost. Each successive round is reviewed by people who are increasingly tired of the reference, so the bar drifts: things that would have been rejected at round one get waved through at round three to make the process stop. That is how a mediocre specification reaches production.

Judgement: Price a revision as a full rebuild plus the loss of the reserved build slot plus any tooling it touches, because the alteration itself is almost never the expensive part.

Change Class Against Calendar Impact

Not all revisions cost the same, and sorting them by class before the mark-up is issued is the single most useful habit in revision management. The class depends on how much of the piece has to be made again and whether anything outside the sewing room has to be re-qualified.

Two classes deserve a warning. Cloth substitution looks like a small change and is not, because the finish has to be re-qualified: water resistance of the new cloth is checked against AATCC 127, and rubbing fastness of any printed mark against AATCC 8. Structural alteration almost always means the bulk slot moves.

Packing is the forgotten class. A revision that changes carton size or inner packing invalidates the transport qualification, and a parcel sequence run to ISTA 3A has to be repeated on the new pack. It is cheap to plan for and expensive to discover after booking.

Spec rule: Classify every revision before issuing the mark-up — colour, geometry, grid, hardware, cloth, structure or packing — because the class, not the size of the alteration, decides whether the bulk slot survives.

Freeze Rules: What Closes at Each Gate

A freeze is a dated statement that a particular set of decisions is closed. Programmes that never say what is frozen drift, because every subsequent conversation reopens everything. Four freezes cover most of what a bag programme needs, and each one should be written with a date and a revision number.

The specification freeze closes the drawing, the dimension list and the material call-out. It should sit at the point where the first round is approved, because that is the first moment at which the buyer has seen a physical piece and confirmed it matches what was intended. Anything raised after that is a change request against a frozen specification.

The component freeze closes the hardware, webbing, thread and closure schedule. It can sit slightly later than the specification freeze, because a buyer may want to swap a buckle after seeing the piece — but it must close before bulk cutting, since a component arriving late stops a line rather than a sample.

The process freeze closes how the piece is made: stitch density, seam construction, reinforcement method, and any jig in use. It matters most where the product carries load, and it is the freeze most often broken quietly when a different operator or a different machine is used.

The packing freeze closes carton, inner packing and marking. It sits latest of the four and is the one most programmes forget, which is why packing revisions arrive after the vessel is booked.

Selection rule: Record four dated freezes — specification at first-round approval, components before bulk cutting, process at prototype sign-off and packing before freight is booked — because a freeze without a date is a preference, not a control.

How Many Rounds a Modular Construction Really Needs

A modular shell with an attachment grid carries more interfaces than a plain bag, so it earns a different round budget. Every interface — a pouch mounting face, a webbing channel, a removable panel — is a place where two parts have to agree, and each agreement has to be checked on a physical piece rather than on a drawing.

Round one on such a build is almost always about geometry: whether the grid spacing puts pouches where the designer intended, whether panels sit flat, whether the shell holds its shape when loaded. Round two is about the modules themselves — whether they mount easily, whether they stay put, whether removal is practical with the bag on. Those are genuinely different questions and both deserve a build.

A third round on a modular reference is usually a signal rather than a need. Most often it means the specification was incomplete at the start: a dimension was never stated, a module interface was never drawn, or an acceptance test was never defined. Adding a third build treats the symptom; completing the specification treats the cause.

The economical response is to spend the round-one preparation time on the interface drawing rather than on the aesthetics. A grid with stated spacing, stated webbing width and a stated module envelope answers most round-two questions before the first piece is cut.

Bottom line: Budget two rounds for a modular construction — one for shell geometry and one for module fit — and treat a third as evidence that the interface specification was incomplete rather than as a routine step.

Running Revisions Without Losing the Bulk Slot

Revision and bulk scheduling are the same problem viewed from two ends, and the programmes that handle them well do one thing differently: they decide the bulk slot question at the moment the mark-up is issued rather than at the moment the revised piece is reviewed. That single habit converts an open risk into a dated decision.

Three tactics carry most of the benefit. Run parallel paths wherever the revision allows — a colourway correction can proceed while a hardware question is still open, because they touch different parts of the build. Pre-book a provisional bulk slot against the earliest plausible approval date, so the slot exists even if it later moves. And set a decision deadline for every mark-up, after which the round is closed on the evidence available rather than left open.

Communication discipline matters as much as scheduling. One named owner on the buyer's side, one channel, and a written mark-up with numbered points prevent the slow spread of half-agreed changes that arrives as six separate messages over a fortnight and lands as a rebuild nobody planned.

Where a revision is unavoidable and large, say so early. A structural change declared while the sample is being built costs one cycle; the same change declared after bulk is cut costs material, labour and the slot — and at a 500-unit floor the material is already bought. Programmes that declare late changes early almost always keep their original bulk booking; programmes that wait for the review to conclude almost never do, because by then the next slot has been taken by somebody else.

Decide the bulk slot at mark-up issue rather than at review, keep independent corrections running in parallel, and close every round on a stated deadline, since an open round consumes more calendar than a refused one ever will.

Gates, Records and the Production Facts Behind the Calendar

Programme work is coordinated through a 4,950 m² floor verified by SGS on which 149 machines sit across 7 production lines, run by 137 people at 200,000 units monthly. Experience of bag production in the organisation dates to 2004 and the business was registered in 2014. Quotations arrive within 24-48 hours. Prototypes are built in 6-10 working days and in 12-15 where the construction is complex; bulk then occupies 35-50 days from a 500-unit minimum per reference. Release is governed by an AQL 2.5 attribute plan — critical zero, major 2.5, minor 4.0 — taken from ISO 2859-1 at general level II, sold FOB Xiamen on T/T 30/70, with sea transit occupying 25-35 days, air 5-8 and courier 3-5.

The record for one reference should hold the mark-up for each round with its issue date, the criteria the round was judged against, the decision with the date and the name, and the photographs or measurements supporting it. Four items per round; where a round produced a deviation note, that note is attached rather than summarised.

Revision history is also what protects the approval later. When a production piece differs from an earlier approved sample, the question is always whether the difference was agreed — and the only answer that settles it is a dated mark-up plus a dated decision.

Revision control connects to the rest of the programme: a MOLLE attachment grid sets the interface dimensions that most round-two comments concern, the custom modular backpack route fixes the gate sequence, sampling fees per round are settled through the terms described in the sampling fee and refund guide, and the whole sequence is scheduled through the prototype and revision services desk. Round structure for a specific reference is best agreed at the enquiry desk before the first build.

Keeping four documents per round — dated mark-up, stated criteria, dated decision and supporting measurements — and recording the deviation notes in full is what lets a revision history settle a later dispute about what was approved.