Home › Field notes › Lead Time Buffer Planning for Launches: Backward Scheduling That Holds

Modular Bag Pro · 2026-10-11 · 2,574 words · 12 min read

Lead time buffer planning for launches means deriving every milestone backwards from a fixed launch date and placing buffer between stages rather than inside them, because a buffer folded into a stage estimate is consumed silently and leaves no slack to recover with. The chain runs quotation in 24-48 hours, sampling in 6-10 working days and 12-15 for a complex build, bulk in 35-50 days, and sea freight in 25-35 days, so a launch date set without working backwards from those figures is a wish rather than a plan. The boundary is that planning figures are planning figures: quote the upper end of each range, and treat any node you do not control as the one that will slip.

Backward Scheduling: The Launch Date Sets Every Earlier Date

Forward planning asks how long each step takes and adds them up, which produces a date nobody is accountable for. Backward planning starts from the one date that is genuinely fixed — the launch, the retail reset, the trade show, the catalogue drop — and works backwards, subtracting each stage and each buffer in turn. The result is not a prediction; it is a set of deadlines, and a deadline missed by three days is visible on day three instead of on launch day.

The arithmetic has one rule: plan at the upper end of every quoted range. Sampling is quoted at 6-10 working days and 12-15 for a complex build, so the plan uses 10 or 15. Bulk is quoted at 35-50 days, so the plan uses 50. Sea transit runs 25-35 days, so the plan uses 35. The spread between the low and high figure is not padding — it is the measured uncertainty of the process, and spending it in the plan is what converts a range into a date.

Read downwards, that table is the plan. Read upwards, it is a set of commitments: the quotation has to be requested roughly five months before launch for a sea-shipped programme, and a launch date announced before that conversation has happened has already consumed the schedule.

Selection rule: Build the calendar backwards from the launch date using the upper figure of every quoted range, and record the spread between low and high as the buffer for that stage rather than as optimism.

Where Buffers Belong: Between Stages, Not Inside Them

A buffer hidden inside a stage estimate disappears the moment the stage runs long. If bulk is planned at 42 days because 35 plus a 7-day cushion felt safer than 50, then a bulk run that genuinely takes 50 days does not consume a buffer — it simply overruns, and nobody sees it until the packing week. A buffer held between stages is a separate line with its own owner and its own visibility, which is the entire point of having one.

Between-stage buffers also do something else important: they absorb handover losses. The gap between production finishing and the container being booked is not production time and is not freight time; it is document time, and it is where a commercial invoice that does not match a packing list quietly costs a week. Holding a named buffer at each handover converts an invisible loss into a visible one.

Three buffers earn their place in almost every launch plan. One between PP approval and production start, which absorbs late artwork changes discovered on the physical sample. One between production finish and vessel departure, which absorbs a booking rollover or a document correction. One between arrival and the launch date, which absorbs customs entry questions and inland delivery. Each is entered as a line, not folded into an estimate.

Seasonality earns a separate line rather than a general cushion. Production calendars in this trade move with public holiday periods and with the pre-peak rush ahead of them, and a launch timed to arrive inside a peak window competes for bookings and for line slots with every other programme on the water. The quoted ranges do not contain that contention, so a named seasonal buffer is added to the production and booking lines during those months instead of being discovered as a surprise in week six.

Verdict: Hold buffers as separate dated lines at each handover — approval to production, production to vessel, arrival to launch — and never blend a cushion into a stage duration where it cannot be seen or defended.

The Nodes That Slip Most Often and Why

Slippage is not random. In bag programmes the same four nodes account for most of it, and each slips for a structural reason rather than through carelessness. Naming the reason is what makes the buffer appropriate rather than arbitrary.

Sampling slips because it is the stage with the most unknown content: a build that has never been made needs pattern work, and a complex build needs 12-15 working days rather than 6-10. Approval slips because it is the only stage owned by the buyer rather than by the production side, and buyer-side review rarely respects its own deadline. Material procurement slips because a mill's lead time is not the same as a cutting room's, and a fabric confirmed late pushes everything after it. Booking and customs slip because both depend on third parties whose calendars are not ours.

The common thread is ownership. Three of the four nodes are either owned by the buyer or by a third party, and the reasonable planning response is not to hope they hold but to size the buffer against the party with the least incentive to be on time.

Bottom line: Size each buffer against who owns the node rather than against how long it should take, and put the largest buffers on buyer-side approval and third-party booking.

Freight Mode as a Buffer Decision, Not a Cost Line

Freight is usually chosen on rate, which is the wrong axis when a launch date is fixed. Sea runs 25-35 days, air runs 5-8 days and express runs 3-5 days, and the difference between the sea and air figures is roughly a month of calendar. That month is the largest single buffer available in the whole chain, and buying it at the end of the plan costs far less than losing the launch slot.

The right way to use it is as a decision with a trigger, not as a preference stated at the start. Plan on sea. Set a decision date — the point at which switching to air still saves the launch — and name the person who makes the call. If production finishes on time, sea holds and the budget is intact. If production overruns past the trigger date, air is authorised automatically rather than debated while the goods sit at the port.

Partial air is the option most programmes forget. Air-freighting 20 percent of the order to open the launch on time, with the remainder following by sea, costs a fraction of flying the whole lot and protects the date. It works best where the launch quantity is front-loaded into a few key accounts rather than spread evenly.

Takeaway: Plan on sea freight, set a named trigger date for switching to air, and treat partial air-freighting as the standard compromise rather than flying the full order or missing the date.

Sampling and Approval: The Gates That Consume the Calendar

Sampling and approval together consume more calendar time than their quoted durations suggest, because each can loop. A sample that needs a second revision does not add one sampling cycle; it adds sampling plus a second approval cycle, and each loop is 6-10 working days on routine work or 12-15 on a complex build plus however long the buyer takes to look at it.

Two disciplines limit the loops. The first is to review the sample against the written specification rather than against an impression, so comments are specific and actionable instead of subjective. The second is to cap the revision rounds: two rounds is normal, and a third usually indicates the specification was never agreed rather than that the sample is wrong.

The pre-production sample is the gate that matters most, because it is what bulk is built against. Everything approved at that gate — construction, materials, colour, label content including origin and care symbols — enters production as the reference, and anything changed after it is a change to a running order rather than an edit. Origin wording in particular belongs in that freeze, because a change after artwork release reprints the whole batch.

Judgement: Treat the PP sample as the hard freeze on construction, materials and all label content, and resolve every open question before it, because a change afterwards costs more in calendar time than the change is worth.

Recovery Order: What to Cut First When a Node Is Missed

When a node is missed, the instinct is to compress everything downstream, which spreads the damage across the whole chain instead of absorbing it in one place. A better approach is a fixed recovery order: decide in advance which lever is pulled first, so the decision is made in a minute rather than argued for a day.

The order that works in practice is: first, spend the between-stage buffers, starting with the one closest to the missed node. Second, switch freight mode against the trigger date — air at 5-8 days or express at 3-5 days buys back the largest single block of calendar time in the plan. Third, reduce scope: ship the launch quantity first and the replenishment later, or drop a colourway from the opening shipment and follow with it. Fourth, move the launch date, which is a business decision and should only be reached after the first three have been costed.

What should not be cut is inspection. Release sampling follows ISO 2859-1 at level II with a 2.5 AQL on majors and a 4.0 on minors and no acceptance number on criticals, and skipping it to save four days converts a late launch into a defective one. The inspection window is days; the recall is months.

Recovery is cheaper when it is pre-authorised. A plan that names who may switch freight mode, up to what cost, and who may approve a split shipment, recovers in hours. A plan that requires a budget meeting to release an air freight differential loses most of the time it was trying to buy.

Spec rule: Fix the recovery order in the plan before launch — buffers first, then freight mode, then scope, then date — and pre-authorise the freight switch to a named person with a cost ceiling.

One Master Calendar With Named Owners

A launch plan spread across email threads, a sampling schedule and a freight booking sheet is not a plan, because nobody can see whether it is slipping. One document, one column of dates, one named owner per line, and one review cadence is the whole discipline. The document does not have to be elaborate; it has to be single.

Each line carries four fields: the milestone, the date, the owner and the dependency. The dependency field is the one that gets skipped and the one that matters, because a milestone that depends on a buyer-side approval has a different risk profile from one that depends on a cutting room. Where a line has no owner, it has no date in practice.

Review cadence should match the risk. Weekly is enough until sampling starts; during sampling and PP approval, twice weekly, because that is where loops happen; during production, weekly against a production snapshot; from vessel booking to arrival, twice weekly, because that is where third-party calendars bite.

Two dates on that document carry more weight than the rest and belong on page one rather than buried inside a stage. The first is the freight-mode trigger: the last day on which a switch from sea to air still protects the launch. The second is the artwork freeze: the day label content, carton marks and document templates stop changing. Every recovery decision eventually turns on one of those two.

Programme work is coordinated through a 4,950 m² SGS-verified site: 7 lines, 149 machines and 137 people, with monthly output near 200,000 units, and bag production experience in the founding team reaching back to 2004 at an operation set up in 2014. Capacity is allocated against that calendar, which is why a slot requested after the schedule is set competes with programmes already on the lines rather than being absorbed invisibly.

Commercial Terms and Capacity Behind a Launch Calendar

The commercial sequence sets the front of the calendar. A quotation returns in 24-48 hours against a specification that is complete enough to cost; a vague enquiry does not return faster than a complete one, it returns vaguer. Terms are T/T 30/70, with 30 percent releasing materials and 70 percent before shipment on FOB Xiamen terms, and the deposit is what starts the material procurement line in the backward schedule rather than the purchase order date.

Capacity interacts with the calendar in one specific way: a slot is held by deposit and confirmed specification, not by conversation. Bulk runs 35-50 days and the line has to be allocated before that window opens, so the practical deadline for locking a launch slot is the point at which the schedule reaches the production line — roughly 108 days before launch on the table above, and earlier in a peak season.

Documentation closes the back of the calendar. Carton cube is planned at roughly 28 CBM to a 20GP and 68 CBM to a 40HQ, the packing list is generated from the same carton count as the marks on the boxes, and the documents are reconciled before loading rather than during transit. Market-specific import and entry requirements are published by the U.S. Department of Commerce, and parcel-transit performance of packed cartons is characterised through the ISTA 3A procedure — both worth consulting while the calendar is being built rather than after it slips.

Programmes with a hard external launch date benefit most from a planned custom modular backpack programme rather than a spot order, because the slot, the material bookings and the document cycle are then reserved against a date. Landed-cost consequences of switching freight mode mid-plan are set out in the note on calculating landed cost for custom modular bags, and the sampling sequence that feeds the calendar is described under our services page.