Home › Field notes › Capacity Planning for Bag Orders: Line Allocation and Rush Windows

Modular Bag Pro · 2026-10-11 · 3,927 words · 18 min read

Capacity planning for a bag order means reserving named production lines, input materials and inspection slots before the purchase order arrives, rather than accepting a date and hoping the floor absorbs the volume. A 4,950 m² SGS-verified production floor running 7 production lines with 149 machines and 137 people plans monthly output around 200,000 units, and capacity is consumed in line-days: an order starting at MOQ 500 occupies a defined block of them, sampling takes 6-10 working days, and mass production takes 35-50 days once approvals and inputs close. Whether a rush request can be met depends on what the relevant line is doing at that hour — its changeover state, its input position and its inspection queue — not on how eager anyone is to agree. Every figure quoted while a sample is still under review is indicative only, rests on FOB Xiamen terms, and covers civilian load carriage such as hiking, commuting, worksite tools and first aid.

What Capacity Means When a Bag Order Is Actually Booked

A monthly output figure is a planning abstraction. When a buyer reads that a site is scheduled around 200,000 units per month, the useful question is what that number means for one purchase order, and the honest answer is measured in line-days rather than in pieces. A line-day is one production line working one shift at the rate it can actually hold for a given reference. An order is genuinely booked when its line-days, its input lead time and its inspection slot have all been placed on a calendar; it remains at risk when only the ship date has been agreed in an email.

Four consumers of capacity sit behind every bag order, and only one of them involves a sewing machine. Cutting and preparation hold space and labour before a needle moves. Sewing holds the allocated line or lines for the duration of the run. Hardware and trim kitting holds bench space, and it is frequently the quiet constraint because buckles, webbing, zips and labelled trims arrive on their own clocks. Inspection holds a slot at the end. A buyer who plans around sewing alone will book a date the other three cannot support.

MOQ 500 exists for capacity reasons as much as commercial ones. Set-up, first-piece approval, pattern confirmation and changeover consume roughly the same calendar whether the run is 200 pieces or 900, so a small quantity spends a disproportionate share of its line-day budget on work that produces nothing shippable. Above the minimum the fixed portion spreads, the marginal time cost per unit falls, and the line reaches its plateau faster in relative terms.

Sampling is the first real draw on that budget. A standard construction returns in 6-10 working days, extending to 12-15 where a moulded part, an attachment interface or a new jig has to be prepared. Those days occupy the sample room, the pattern bench and machine time that a running order would otherwise use, which is why a rush request raised in the middle of sampling competes directly with the sample the buyer is waiting for.

The practical consequence is simple to state and easy to ignore. Two enquiries received on the same day can be given different honest answers if one arrives with a frozen tech pack and approved components while the other is still changing hardware. Capacity is not a queue; it is a schedule of named resources, and the buyer who supplies decisions early buys a better position in it.

Takeaway: Book a bag order against line-days, input lead time and an inspection slot rather than against a promised date, because a 500-piece reference consumes fixed set-up capacity of roughly the same size whether 200 or 900 units eventually run.

Line Allocation: Matching a Reference to the Right Production Line

Seven production lines are not seven identical lanes. Each carries a different mix of the 149 machines on the floor, a different operator skill profile, and a different history of what it ran most recently. Allocation is therefore a matching exercise: the construction, the hardware content, the seam count and the finishing operations decide which line can hold the target daily rate without a long re-learning curve, and a poor match costs more days than most buyers ever see itemised.

A hardware-heavy build belongs on a line whose benches are already set for riveting, bar-tacking and buckle assembly. A panel-heavy build with long uninterrupted seam runs belongs where the machine mix supports continuous sewing without a mid-cycle bench move. A reference carrying an attachment grid behaves differently again, because webbing consumption and stitch island count change both the operation sequence and the inspection burden; the process notes attached to the MOLLE attachment grid specification show how much interface detail can move the operation time of an otherwise ordinary body.

Recent history matters as much as equipment. A line that finished a comparable construction last week re-threads faster, sets its jigs from memory and reaches plateau sooner than a line meeting the reference for the first time. That difference is why two nominally equivalent lines can be given different start-up expectations for the same drawing, and why moving a repeat order to a different line is never a neutral administrative act.

Allocation also determines the failure surface. A reference running on one line has one point of failure: if that line stops for a machine, an input or an operator absence, the order stops. Splitting across two lines doubles the supervision and inspection burden but halves the consequence of a single stoppage — provided both lines are released against the same approved reference sample and the same drawing revision. Without that condition, a split produces two populations of units that a receiver may treat as two different products.

Buyers can influence allocation by being specific early. Stating the dominant operation, the hardware family, the expected seam count and the packing configuration lets a planner put the reference on the right line the first time, instead of discovering at first-piece stage that the allocated line cannot hold the rate and re-planning mid-run.

Spec rule: Allocate a reference to the line whose machine mix and recent history match its heaviest operation, and accept a two-line split only when both lines are released against one approved sample revision and one drawing revision.

Daily Output Ramp: From First Piece to Plateau to Close-Out

Daily output is not constant across a run. It starts low, climbs through a learning stage, plateaus, and then falls again at the close as the line is stripped for the next reference. Buyers who plan against the plateau rate as though it applied from hour one under-estimate the calendar with remarkable consistency, and the error compounds when the same assumption is applied to a second colourway.

The climb has identifiable stages. First pieces are produced slowly while operators re-read the construction and inspectors confirm the build against the approved reference. Work-in-progress buffers then fill between operations, the line finds its rhythm, and the count approaches the level the operation was designed to hold. Only after that does a daily figure mean anything as a planning input.

The plateau itself is set by the slowest operation, not the fastest. Where a bench operation for hardware takes longer per unit than the sewing ahead of it, units queue between the two and the daily rate becomes the bench rate. Adding operatives upstream of that constraint increases the queue without increasing output; removing or rebalancing the constraint is the intervention that actually moves the number.

The close-out decline is the stage most often omitted from planning. As the run finishes, the remaining sizes, colourways and packing variants are completed in smaller batches, changeovers multiply, and the effective daily rate drops well below plateau. Booking a vessel on plateau arithmetic for the final portion of quantity is a dependable way to miss the cut-off.

Reading the ramp correctly also changes how progress reports should be interpreted. A daily count reported during the climb is not evidence of a problem; a daily count that fails to reach plateau by the end of the second stage is. Buyers who ask for the stage alongside the number get a far more useful picture than buyers who ask for the number alone.

Bottom line: Plan a run in five stages rather than at one flat rate, and never book transport against the plateau figure for the final 12-18% of quantity, where colourway and packing changeovers pull the effective daily rate down.

What an Interrupt Order Costs Inside a Running Line

An interrupt is not free even when it is small. The visible cost is the stoppage; the real cost is the restart. A running line holds a set of conditions — threaded machines, set jigs, positioned work-in-progress, operators in rhythm and an inspector calibrated to the current reference. Pulling a different reference through the middle of that destroys the set, and the set then has to be rebuilt twice: once for the interruption and again when the original run resumes.

The rebuild items are concrete and countable. Machines are re-threaded with a different thread and often a different needle size. Jigs, guides and templates are swapped and then swapped back. Partially finished units are parked, and parked work-in-progress is both a space cost and a damage risk, because stacked sub-assemblies pick up soil, creasing and hardware marking while they wait. Operators are re-briefed on a construction they have not seen this week, and a first-piece check is repeated for the interrupting reference and then repeated again for the original.

The practical penalty usually lands between half a day and two full days of lost plateau output for the displaced order, and the position inside that range is decided by material commonality. Where the two references share a base fabric, thread specification and hardware family, the penalty sits at the low end. Where they do not, it sits at the high end, and the displaced order additionally loses its inspection slot because the inspection queue does not re-sequence as easily as the line does.

A second-order effect sits in materials rather than in labour. An interruption can consume trim that was kitted for the original order, and re-kitting after the fact is slower than the original kit because the bench has moved on to other work. Where the interrupting reference needs an input that is not yet on hand, the line stops for the input rather than for the changeover, and that stop is outside the control of anyone on the floor.

The defence is a written interrupt rule agreed before the run starts: what size of insertion will be accepted, from which references, at what notice, and with what compensation to the displaced order. Agreed in advance, an interrupt is a priced commercial decision. Agreed under pressure, it becomes a dispute about who caused the delay.

Judgement: Accept an interrupt only when the two references share thread, needle and hardware family, because a shared-material insertion costs roughly half a day of plateau output while a full changeover, re-start and re-inspection costs up to two days plus a lost inspection slot.

Peak Season Windows and the Three Scarce Resources They Compete For

Bag demand is seasonal at the retail end, and the season reaches the production end earlier than most buyers expect. A shelf date set by a campus cycle, a holiday programme or a corporate rollout translates backwards into a production window, and that production window then competes with every other brand targeting the same week on the same shelves.

The competition is for three scarce resources and nothing else: line-days, input positions and inspection slots. Two buyers can each be told their order fits, because both were quoted against a snapshot of the calendar taken before the other one booked. That mechanism, rather than any refusal, explains most peak-season date failures — the schedule was honest when written and stale by the time it mattered.

Transit is the block that cannot be compressed once booked. Ocean movement at 25-35 days forces the production close-out backwards by roughly a month relative to the date goods are actually needed, and any slippage inside that month has to be recovered by air at 5-8 days or courier at 3-5 days, each with its own cost profile, documentation burden and volumetric consequences. Peak season is therefore a freight planning problem at least as much as a sewing problem, and mode choice belongs in the same conversation as line allocation, as set out in the comparison of sea, air and express freight modes.

Input positions fail earlier and more quietly than lines do. A specific coated fabric, a plated buckle in a specified finish, or a branded zipper family can have a procurement clock longer than the production run itself. Buyers who reserve line-days but leave hardware to be confirmed after sample approval routinely discover in week three that the line is free and the bench is empty.

The workable response is to move the reservation point earlier. A forecast with a tolerable band, a component approval completed before the seasonal rush, and a standing slot for a repeat reference all convert an optimistic snapshot into a held position. Buyers running a defined range can also shorten the conversation by freezing the platform once, as described for the modular backpack platform family, so seasonal volume becomes a quantity decision rather than a development decision.

Verdict: Treat a peak season delivery as three separate reservations — line-days, component positions and an inspection slot — because a quote covering only one of them is a snapshot, not a booking, and ocean transit at 25-35 days leaves no room to recover lost production time.

Why 'Can You Rush It' Is a Line Question, Not a Willingness Question

The question is asked in every programme and answered badly in most of them. Willingness is uniform; capacity is not. A line sitting at plateau on a reference that shares thread, needle and hardware with the urgent request can absorb a modest insertion at modest cost. The same line, one day into a changeover for a completely different construction, cannot absorb anything without paying the full restart price described earlier. The answer therefore belongs to the planner holding the schedule, not to the person answering the email.

Three genuine mechanisms exist for recovering time, and each has a ceiling. Adding a second line to the same reference recovers real days but requires a second approved reference sample, duplicated jigs and duplicated inspection cover; without those it creates revision drift instead of output. Overtime and weekend working add hours at the plateau rate, bounded by sustainable working patterns and by whether inspection capacity extends to cover the extra output. Re-sequencing the queue recovers nothing on its own — it transfers time from one order to another, which is a commercial decision about whose date moves.

A fourth mechanism is often proposed and rarely delivers: compressing the input clock. Ordering fabric and hardware after sample approval rather than in parallel with it is the single largest avoidable delay in most programmes, and no amount of line pressure recovers it. Buyers who approve components early buy more schedule relief than buyers who negotiate rush language.

The honest answer to a rush request is therefore a short schedule statement rather than a yes or a no: what the line is running, what shares material with the request, what inspection cover exists, and what the two or three recovery mechanisms would each yield. Buyers who receive that statement can make a real decision. Buyers who receive an unqualified yes discover the mechanism later, usually at the shipping stage.

Selection rule: Ask which mechanism will be used to recover the days — second line, extended hours or queue re-sequencing — and accept the request only when that mechanism's ceiling exceeds the gap, because willingness alone produces no line-days.

Capacity Arithmetic: Working Backwards From a Delivery Window

Backwards planning is the only reliable way to test whether a delivery window is real. The method is to start at the date goods must be available and subtract each unavoidable block in turn, then compare the resulting start date with today. Where the start date sits in the past, the programme is already late, and the only question left is which block will be compressed and at what cost.

Two adjustments make the arithmetic honest. First, add the production ramp and close-out stages as calendar rather than as units, because a run quoted at 35-50 days on plateau arithmetic does not include the slow ends. Second, treat the component clock as a parallel path that must start no later than sampling sign-off rather than as a sequential step after it.

The arithmetic also exposes where the risk actually sits. In most programmes the largest controllable delay is not the run; it is the interval between sample dispatch and buyer approval, followed by component procurement. A buyer who compresses those two gains more schedule than a buyer who negotiates the production number down by a week.

Where the window genuinely cannot be met on ocean transit, the decision is a mode decision with a cost consequence rather than a production decision. Compressing production to protect an ocean booking is usually cheaper than compressing production and then paying for air, but only if the compression stays inside the ceilings described in the previous section.

Programme Terms and the Production Base Behind the Schedule

Our production team sequences bag orders against line state, component position and inspection cover rather than against enquiry date, and the founder has worked in bag production since 2004 with the company established in 2014. Assembly is distributed across a 4,950 m² SGS-verified production floor staffed by 137 people, planned around 7 production lines and 149 machines, with monthly output scheduled near 200,000 units. Programme flow runs sampling, PP sample, AQL 2.5 inspection and shipment, with final inspection sampling drawn from the system published by ISO 2859-1.

Commercial terms stay constant whatever the capacity conversation. The order minimum is 500 units per reference, sampling occupies 6-10 working days and extends to 12-15 on a demanding construction, and mass production takes 35-50 days once approvals and inputs have closed. Settlement runs on T/T split 30/70, and the agreed delivery term is FOB Xiamen, so everything beyond the ship's rail at the load port — ocean movement, insurance, destination handling and clearance — belongs to the buyer's own landed cost model rather than to the quotation.

Capacity commitments are only meaningful when they are tied to those gates. A date quoted before sample approval assumes approval arrives on schedule; a date quoted before component approval assumes the bench will be full when the line is free. Both assumptions are reasonable and both should be written down, because a schedule built on an unstated assumption is the source of nearly every dispute about who caused a delay.

Documentation supports the schedule in both directions. Export-side paperwork follows the guidance material published by U.S. Department of Commerce for buyers importing into that market, while ISO 9001 describes a management system framework and not the conformity of any individual shipment. Nothing here applies to weapon carriage, ballistic protection or any defence certification claim; the scope is civilian load carriage only.

A Capacity Planning Checklist to Send With the Enquiry

Most capacity disputes begin as an enquiry that did not say enough. A short, structured enquiry lets a planner reserve real resources instead of estimating them, and it takes one page. The items below are the ones that change the answer when they are missing.

• Reference quantity split by colourway and by packing variant, with the minimum confirmed as 500 units per reference.
• The dominant operation in the construction: continuous seaming, hardware assembly, or interface work such as an attachment grid.
• The approved component list, or a clear statement of which components are still open, because open components start their own clock.
• Target delivery window at destination, plus whether the window is a shelf date or a port arrival date.
• Preferred transport mode and whether the programme can tolerate ocean transit of 25-35 days or requires air at 5-8 days.
• Whether a second production line may be used, and whether a second approved reference sample can be released for it.
• Interrupt policy: what insertion size the buyer expects to request, and what notice they can give.
• Inspection expectation, including whether the buyer will send their own inspector to the AQL 2.5 final inspection.
• Settlement readiness against T/T 30/70, since the deposit releases the component clock.
• Repeat intent, because a reference expected to run again deserves a held slot rather than a fresh negotiation.

Buyers who send that list get a schedule built on named resources. Buyers who send a quantity and a date get a snapshot, and snapshots are the reason peak season dates fail. Where the programme is a development rather than a repeat, the same discipline applies one step earlier, and the custom modular backpack development route sets out how the interface and components are frozen before a schedule is worth discussing.