How Many Sample Rounds a Luxury Bag Needs

Why the number of rounds is a decision, not a fate

Teams treat sampling rounds as something that happens to them. In practice the count is predictable from two questions: how novel is the construction, and how complete is the pack? A novel construction on an incomplete pack will run four or five rounds, and each one is spent discovering information that should have been written down at the start. A proven construction on a complete pack resolves in two, because there is nothing left to discover.

The second variable is decision discipline. A round that produces three contradictory notes from three stakeholders does not resolve anything; it generates the next round. Programmes where one person signs off per round converge far faster than programmes where approval is a committee, regardless of how complex the bag is. The third variable is timing: a sampling window squeezed against a delivery date forces approvals that should have been refusals, and those approvals resurface as bulk tolerance arguments months later, when they are far more expensive to settle.

What actually happens in round one

Round one is a structural question. The pattern is cut, the bag is built, and the questions are about proportion, balance, stand, handle drop and whether the silhouette reads as intended when it is held rather than drawn. Material in round one is usually a stand-in close to the intended specification, because committing the final hide to an unproven pattern wastes expensive leather.

• Does the bag stand correctly, or does it slump at the base?
• Is the handle drop comfortable when the bag is loaded rather than empty?
• Do the panels close without fighting each other at the corners?
• Does the hardware seat flatly, or does the plate rock on unskived leather?
• Is the internal volume usable, and do the pockets sit where a hand expects them?

What round two is for

Round two is the confirmation round, and it is where the final material, the final colour and the final finishing are used. This is the round that becomes the approved reference if it passes. Colour is measured against the signed panel under D65, stitch density is checked at 8 stitches per inch on visible calfskin, edge build is compared to the reference chip, and hardware is cycled and pull-tested.

The discipline that makes round two work is that it should only be answering questions raised in round one. If round two raises new structural questions, something went wrong in the pattern correction, and a third round is already likely. We record every correction from round one and confirm each one on the round two piece before anything else is assessed.

When a third round is justified

Three situations genuinely justify a third round. The first is a material that behaves unexpectedly: an exotic skin whose scale pattern changes the way a panel folds, or a vegetable-tanned leather whose temper shifts during finishing. The second is hardware that has to be re-tooled because the engagement is wrong, which cannot be assessed until the replacement arrives. The third is a client-side design change, which is legitimate but should be acknowledged as a fresh start rather than counted as a sampling failure.

Rounds needed by material

Material drives the count more than silhouette does. A plain calfskin behaves predictably and can be committed to round two with confidence. An exotic skin cannot: the pattern has to be placed against the scale, the yield has to be calculated, and a panel that looked right in calfskin may fold badly in caiman. Vegetable-tanned leather introduces its own variable, because it darkens through finishing and the approved colour is a moving target until the process is fixed.

Where extra rounds come from

Four causes account for most programmes that run past three rounds. The first is an incomplete pack, where substance, tolerance or construction sequence was never written. The second is dispersed approval, where three stakeholders give three sets of notes and no single person resolves the conflict. The third is a change of intent mid-process, where the brief moves between rounds. The fourth is a material substitution nobody agreed to.

1. Write the pack completely before round one: substance, tolerance, construction sequence, stitch density and edge build.
2. Name one approver per round, and require that person to consolidate every other stakeholder's notes into one list.
3. Freeze the brief between rounds; treat any change of intent as a new round with a new date rather than as a note.
4. Confirm the material in writing before round two cuts, including the dye lot reference and the tannery batch.
5. Photograph and measure the round one piece before sending it, so the corrections are evidence rather than memory.

What a pre-production piece adds

The final build before bulk is not another design round. A pre-production piece is made on the production line, with production tooling, production jigs and the final material, by the people who will make the order rather than by the sampling bench. Its purpose is to prove that the line can repeat the approved sample, and it frequently reveals things a sampling round cannot: a jig that needs adjusting, a station sequence that slows the build, or a component that behaves differently at speed.

We treat it as a gate rather than a round, and it is inspected at the same level as any production lot. If it passes, it is signed, photographed and held as the reference for bulk. If it fails, the line is corrected rather than the design, and that distinction is worth keeping clear, because a pre-production failure is information about the line rather than about the bag.

Sampling a repeat versus a new style

A repeat order of a style with a signed reference in the file needs far less development than a new one. Where the construction, the material and the hardware are unchanged, a repeat can often be confirmed with a single piece rather than a full round, because the questions have already been answered and the reference already exists. What still needs confirming is material: a new dye lot, a new tannery batch, or a new plating bath all reintroduce the variables that round two was designed to close.

A colourway extension sits between the two. The construction is proven, so the structural questions are already closed, but the new shade has to be measured against its own signed panel and the hardware finish has to be matched to it. In practice a colourway extension costs one round rather than two, provided the new shade is reserved against a single dye lot before cutting starts.

How hardware lead time shapes the schedule

On any programme with bespoke or re-tooled hardware, the schedule is set by the hardware rather than by the bench. Sampling a closure can be completed in days; tooling a new one takes weeks, and no amount of urgency at the bench compresses that. The practical consequence is that hardware should be ordered before the first sampling round, using a provisional specification, so that the real component arrives in time for round two.

Where that sequencing is missed, programmes end up sampling with a stand-in closure and discovering the engagement problem in round three, after the tooling window has closed. We flag it early and ask brands to approve a provisional hardware specification at the same moment they approve the pack, precisely to avoid that outcome.

The cost of an extra round

An extra round costs two things: working days and material. Six to ten working days on a standard build, twelve to fifteen on a complex one, and that delay sits directly in front of the bulk window, because bulk cannot start until a sample is signed. Material cost is smaller but real, particularly on exotic skins where a single mis-cut panel is not recoverable.

Correcting one panel once a round has been built costs around USD 4.80 in bench time before the re-cut and the re-assembly are counted. Multiply that across the corrections in a round and the argument for a complete pack becomes arithmetic rather than opinion.

How to reduce the count without losing quality

Three changes reliably cut a round without lowering the standard. Put the final hardware into round one rather than a stand-in, so seating and engagement are assessed from the start. Use the intended interlining and reinforcement in round one, because they determine stand and hand more than the outer material does. And approve colour against a signed panel early, so that round two is confirming rather than discovering.

When to stop and approve

Approve when every correction from the previous round has been confirmed closed and no new structural question has appeared. Do not approve because the season is late or because the piece is close enough; close enough in sampling becomes a tolerance argument in bulk, and a tolerance argument in bulk is far more expensive than one more round. Where a deadline genuinely cannot move, the honest answer is to approve the sample and re-specify the questioned detail in writing for bulk, rather than to leave it unspoken and hope the production run resolves it by itself. Once approved, the sample is signed, stored away from light, and becomes the reference that inspection compares against.

Standard rounds, proven construction

2 rounds

Complex or exotic construction

3 to 4 rounds

Working days per round, standard

6 to 10

Working days per round, complex

12 to 15

Round one purpose

Pattern, proportion, stand, hardware seating

Round two purpose

Material, colour, stitch density, edge build, hardware testing

Stitch density confirmed

8 stitches per inch on visible calfskin

Colour confirmation

Delta E 1.0 against the signed panel under D65

Correction cost per panel

Around USD 4.80 in bench time

Approval discipline

One named approver per round, all notes consolidated into one list

Minimum order quantity

500 pieces per style per colourway

Bulk lead time

35 to 50 days from approved sample

Shipment terms

FOB Xiamen

Programme history

Export work since 2014 across 7 production lines

Frequently Asked Questions

People Also Ask