Handbag Maison Project Management Explained: The Numbers, Artefacts and Cadence Behind a Capsule

The Numbers Behind a Three-Style Capsule

A three-style capsule at 500 pcs per style runs on about 180-220 recorded decisions, 6-9 revisions of the technical pack, 2-3 sampling rounds and 35-50 days of bulk, and it is managed through eight standing artefacts rather than through email alone.

Founders often assume programme management means chasing. It does not. It means making every open question visible with an owner and a due date, because the questions that damage a calendar are almost never the ones being discussed. They are the ones sitting quietly in an inbox waiting for somebody to answer them.

Read the third and fourth rows together. Sampling rounds are the visible cost of indecision, while brand-side approvals are the invisible one, and only the second of those is usually under the brand's direct control.

Decision Latency: The Variable Nobody Puts on the Schedule

Decision latency is the gap between a question being asked and an answer arriving, and across a standard capsule it averages 2.4 days per decision. Multiplied across 180-220 decisions, waiting absorbs more calendar time than every development gate combined.

The reason latency is invisible is that it never appears on a production schedule. Schedules record operations; waiting happens between operations, in a place no schedule is looking. A programme can therefore be on time at every gate and still land three weeks late.

• Average latency across a capsule runs about 2.4 days per decision where no turnaround rule is agreed.
• Approvals account for roughly 60% of total waiting time, and most of that sits with the brand rather than the bench.
• A single unanswered material question can hold 500 units of panels uncut for a week without appearing anywhere.
• Latency compounds: a decision waiting on another decision does not start its own clock until the first closes.
• Programmes with a written 24-hour turnaround rule typically close a capsule ten to fourteen days sooner.

The remedy is unglamorous and effective: one shared log, one owner per line, one due date per line, and a review that reads only the overdue rows. Most programmes have a version of this already scattered across email threads. Consolidating it is usually worth more than any schedule optimisation.

The Change-Order Curve, in Multiples

A change costs 1x before the technical pack is frozen, 6-10x after material has been reserved and cut, and 25-40x once units are in assembly. The multiplier is not a penalty; it is the arithmetic of rework, because each stage has more work behind it to undo.

Two consequences follow. First, the least costly possible moment to change your mind is during pack writing, which argues for spending longer there rather than shorter. Second, a spec freeze is not bureaucracy: it is the line where the multiplier steps from 1x to 6-10x, and it deserves a signature.

That is also why the technical pack typically runs to 6-9 revisions. Those revisions are not wasted effort; they are the low-cost end of the curve absorbing changes that would otherwise be paid for at the expensive end.

Eight Artefacts Every Programme Should Hold

Eight documents carry a programme from brief to handover, and a programme is only as manageable as the least maintained of them. None is complex; the discipline is keeping them current rather than complete at the start and stale by week four.

1. Technical pack: dimensioned drawings, materials, hardware, construction notes, held under version control.
2. Decision log: every open question with an owner, a due date and a recorded answer.
3. Material schedule: tannery, lot reference, finish, quantity reserved and arrival date.
4. Component schedule: hardware drawings, plating specification, plating batch and proving result.
5. Sampling record: each round, the correction list issued, and the correction confirmed closed.
6. Inspection plan: AQL 2.5 under ISO 2859-1, the sample size and the accept and reject numbers agreed before bulk.
7. Compliance file: test reports per lot against each destination market, plus permit documents where they apply.
8. Handover pack: packing list, commercial documents, inspection result and the retained reference piece.

The material and component schedules are the two most often skipped, and they are also the two that decide whether a substitution is caught early. When a tannery cannot supply, the schedule is what turns a quiet replacement into a conversation.

Version control deserves a word. A technical pack that has been revised nine times and exists as nine attachments in nine emails is not under control. One file, one version number, one place, and every previous version retrievable.

Who Owns Which Decision

Every recurring decision on a programme has an owner on one side or the other, and ambiguity about which side is the single most common cause of a missed date. The allocation below is the one we work to, and it can be agreed in a single call.

Design intent and geometry

Brand. Signed off at the technical pack, then frozen.

Material selection

Brand chooses; the programme partner confirms availability and yield.

Lot reservation timing

Programme partner, once the brand names the material direction.

Pattern and grading

Programme partner, measured against the approved drawing.

Correction lists after sampling

Programme partner raises; brand confirms closure within 24 hours.

Compliance testing scope

Programme partner books; brand states the destination markets.

Inspection sample plan

Agreed jointly before bulk; neither side changes it afterwards.

Freight and trade terms

Brand selects; FOB Xiamen is the default, with CIF and DDP available.

The third card is where programmes most often fray. A substitution agreed over the telephone is not agreed. It needs the same written confirmation as a specification change, precisely because it is the kind of change that only becomes visible once it is too late to reverse.

The Weekly Rhythm That Keeps a Programme Honest

A capsule is kept honest by a fixed weekly cadence rather than by ad hoc chasing: one written status note, one gate review, one bench walk-through and one overdue-items read. Four touchpoints, the same day each week, is usually enough.

• Monday: a written status note covering progress, open items and anything the brand must decide this week.
• Tuesday: gate review of 45 minutes, reading only the items that block the next stage.
• Wednesday: a bench walk-through with photographs, so geometry questions are settled before cutting.
• Thursday: the overdue-items read, where any line past its due date gets either an answer or an owner change.

Photographs on Wednesday matter more than they sound. Most geometry disputes that reach assembly began as a three-millimetre disagreement nobody could see, and a photograph taken at the bench settles in minutes what a written description argues about for days.

Escalation: The 24 and 48 Hour Rule

Agree two numbers before the programme starts: a 24-hour target for any answer the programme is waiting on, and a 48-hour point at which the question automatically escalates to a named person on each side. Without the second number, the first is decoration.

The third line is the one that makes the rule work. A question that has gone unanswered for four days is usually not a delay; it is a decision nobody wants to make, and treating it as a scope change forces the conversation that chasing never would.

Escalating to a named person rather than a department matters for the same reason. A message addressed to a team is owned by nobody, and it will sit exactly as long as the question was always going to sit.

Risk Register: Twelve Items Worth Tracking

A programme risk register works when it holds twelve items or fewer and each has a named trigger. Longer registers are written once and never read. The eight below account for the great majority of what actually goes wrong.

The value is in the trigger column. A risk without a trigger is an anxiety; a risk with one is an instruction, and it can be acted on by whoever happens to be on the bench that morning.

What a Good Status Report Contains

A useful status report fits on one screen and answers four questions: what moved, what is blocked, what you must decide this week, and whether the date still holds. Anything longer is read by nobody.

1. What moved since the last report, stated as progress against the current gate rather than as activity.
2. What is blocked, with the owner named and the number of days it has been waiting.
3. What the brand must decide this week, with a recommended option and the cost of not deciding.
4. Whether the delivery date still holds, stated plainly, with the reason if it does not.
5. A photograph of the current bench state, because most disputes are geometry and most geometry is visible.

The fourth line is the one reports usually omit, and it is the only line most readers actually want. A report that says a date still holds is useful; one that says it has slipped by four days and why is far more useful, because four days can still be recovered.

How to Close a Programme Properly

Closure is a gate like any other: inspection result recorded, corrections closed and re-checked, documents reconciled, and a retained reference piece set aside for the re-order. Skipping it is why so many second seasons start from scratch.

The retained reference is the item brands most often forget to ask for. It is the physical record of what was approved, and the only thing that settles a colour or finish dispute eighteen months later.

Our programme works to MOQ 500 pcs per style and colourway, sampling at 6-10 working days and bulk at 35-50 days on FOB Xiamen terms. It runs across 7 production lines with about 200,000 pcs per month of capacity and export experience since 2014, delivered through our validated partner factory Quanzhou Yingming Bags, an SGS-validated site.

Frequently Asked Questions

People Also Ask