Handbag Private Labelby QUANZHOU JUNYUAN BAGS Get a quote

How to Communicate Bag Design Changes to a Sample Room

How should bag design changes be sent to a sample room? As one numbered revision list, with one change per line, no more than 12 lines per round, three annotated photographs and a stated dimension for every change. The single most expensive habit in sampling is the follow-up email: changes sent across four separate messages reset the 6-10 working day window three times, because a bench cannot start a round it knows will be interrupted. A private label coordinator should acknowledge a revision list within 24 hours and return the revised sample in 3-6 working days, and the rules below are what make both numbers hold.

How to Send Bag Design Changes to a Sample Room

How should bag design changes be sent to a sample room? As one numbered revision list, with one change per line, no more than 12 lines per round, three annotated photographs and a stated dimension for every change. The single most expensive habit in sampling is the follow-up email: changes sent across four separate messages reset the 6-10 working day window three times, because a bench cannot start a round it knows will be interrupted. A private label coordinator should acknowledge a revision list within 24 hours and return the revised sample in 3-6 working days, and the rules below are what make both numbers hold.

The rules that keep a revision round inside the window

FormatOne numbered list, R01 to R12
One change per lineNever two requests in one line
Round cap12 lines per round
PhotographsThree angles, arrow plus number
AcknowledgementWithin 24 hours
Revision turnaround3-6 working days
Sampling window6-10 working days, 12-15 complex
MOQ500 pcs per style / colourway

One Round, Eleven Changes, Four Emails

A Rotterdam seller launching a 34 cm coated-canvas tote sent the first sample back with a note: the gusset feels shallow and the logo sits too low. Fair comments, and both were correct. What followed was the problem.

Over the next nine days, eleven changes arrived in four separate emails and one voice message. The gusset request changed twice, from 10 cm to 12 cm and then to 11 cm. Two of the eleven changes contradicted each other: the strap was to be lengthened by 30 mm in one message and shortened by 20 mm in another, because two people on the buyer's side were emailing separately. The bench did nothing for six of those nine days, because it had learned that starting meant restarting.

  • Day 1: first note, two changes, acknowledged within 24 hours.
  • Day 3: second email, three more changes, one contradicting the first note.
  • Day 5: voice message, two changes, no written record at all.
  • Day 8: fourth email, four changes, including one that reversed a day-three request.
printed email thread of handbag sample revision requests beside the sample bag they refer to
Eleven changes across four emails and one voice message: the round took nine days before a single stitch was unpicked.

When the changes were finally consolidated into one numbered list of eleven lines, the revised sample came back in four working days. The work itself had never been the problem. The work had been waiting nine days for a document that took forty minutes to write, and the round that should have cost four days cost thirteen.

What That Round Shows About Change Notes

Three rules fall straight out of it, and none of them is about being stricter. They are about making the change list a document rather than a conversation, because a conversation has to be remembered and a document only has to be read.

  • Consolidate before you send. Collect changes internally for 24-48 hours, resolve contradictions inside your own team, then send once. The cost of waiting internally is hours; the cost of sending piecemeal is days.
  • One channel, one owner. Two people emailing the same sample room will eventually send opposite instructions, and the bench cannot adjudicate.
  • Nothing verbal. A voice message that is not written down is a change that will be missed, and the miss is always discovered at the worst moment.

There is a fourth rule that is less obvious and more useful: send the list even when it is short. A round with two changes in a numbered list is processed faster than a round with two changes in prose, because the numbered list tells the bench how many things to expect and lets it confirm each one as it is done.

How to Write a Line That Cannot Be Misread

A change line has four fields. Miss any one of them and the line becomes an interpretation exercise, and interpretation at a bench is where a 3 mm error is born.

  1. A number. R07, so it can be referenced in a reply and ticked off in a photograph later.
  2. A location, named from a datum. Not the strap but the front-left strap end, measured from the top edge of the body.
  3. A current value and a target value. Gusset depth is 100 mm, change to 110 mm. Two numbers remove all ambiguity about direction.
  4. A reference, where one exists: a Pantone TCX code, a swatch number, a hardware code or a photograph number.

Compare two versions of the same request. Make the gusset deeper is a request that will be interpreted, and interpreted differently by a pattern cutter and by a sewer. R07 — gusset depth at the centre of the base: currently 100 mm, change to 110 mm, tolerance plus or minus 2 mm, see photo 2 is a specification that will be executed and then checked.

The same change, written two ways
FieldWeak versionUsable version
NumberNoneR07
LocationThe gussetGusset at the centre of the base
DirectionMake it deeper100 mm to 110 mm
ToleranceNot statedPlus or minus 2 mm
ReferenceNoneAnnotated photograph 2, arrow R07
ExpectationLooks betterRevised sample in 3-6 working days

Which Changes Are Free and Which Reset the Clock

Not every change costs the same, and knowing which is which lets a buyer decide what to ask for now and what to hold for the second order. The dividing line is whether the change touches the pattern.

Colour or hardware swap within the approved range
Free, no pattern change
Branding position moved within 5 mm
Free, jig adjustment only
Dimension change up to 10 mm on one panel
Cheap, one pattern revision
Strap drop change above 30 mm
Moderate, affects wearing and panel balance
Silhouette change or new panel break
Expensive, effectively a new pattern
Closure type change after tooling
Expensive, USD 150-600 for a new die
Material substance change
Resets sampling: 6-10 working days again
Anything after the bulk run starts
Priced per occurrence, plus schedule risk

The last two rows are where programmes go wrong. A substance change is not a revision, it is a new sample, because drape, hand feel and yield all move together and none of them can be predicted from the previous one. And a change after the bulk run has started is priced per occurrence for a reason: it interrupts a line that is balanced across 7 production lines and about 200,000 pcs per month of scheduled work.

How to Annotate a Photograph So It Gets Read

Annotated photographs do more work than prose, and most of them are wasted because they are annotated badly. Three rules make them usable.

  • Three angles per change, minimum: a full view showing where the change is, a mid view, and a close view with something for scale, such as a steel rule or a coin.
  • One arrow, one number, matching the list. An arrow labelled R07 that does not correspond to R07 in the list is worse than no arrow at all.
  • Photograph in daylight or at 5000 K, not under warm shop lighting, because a colour comment sent under warm light will be evaluated under different light on the other side.

A short screen recording beats a photograph for one category of change only: anything that moves. A strap that slips off the shoulder, a drawcord that will not feed, a magnetic closure that pops under load — these are behaviours, and behaviour is easier to show in eight seconds of video than to describe in three sentences. Pair the clip with a written line anyway, so the record exists.

annotated photograph of a handbag sample with numbered arrows and a steel rule for scale
Three angles, one arrow per change, and a steel rule in the close view: the annotation format that gets executed correctly first time.

What a Revision Round Should Contain, End to End

A round is a package, not a message. Send all six parts together and the round moves; send four of them and the round waits for the other two.

  1. The numbered revision list, capped at 12 lines, with current and target values on every dimensional change.
  2. Annotated photographs, three angles per change, arrows numbered to match the list.
  3. A marked-up tech pack page or dimension sheet, so the change lands in the document that will be cut from.
  4. A statement of what is approved, so the bench knows which parts are frozen and must not drift.
  5. A date expectations are due, so the 3-6 working day revision turnaround is measured against something.
  6. One named owner for any follow-up question, with a commitment to answer within 24 hours.

The fourth item is the one most buyers skip and the one that prevents the most damage. Stating what is approved stops a revision from quietly becoming a redesign. Without it, a bench that has been told to change the gusset may also improve the pocket layout, and the next round contains a change nobody asked for.

Where a programme needs more than one round, the sample rounds guide sets out how many is normal for each silhouette type, and the mistakes that add weeks to a launch covers the adjacent traps.

When to Stop Revising and Freeze the Specification

There is a date after which a change costs more than it is worth, and it should be written into the plan at the start rather than discovered through argument. Working backwards from the bulk window of 35-50 days, the practical freeze point is 30 days before the production slot begins, because that is roughly how long a hardware component with custom plating needs to be ordered and received.

  • Freeze dimensions 30 days before the slot. Anything later collides with material purchasing.
  • Freeze hardware at the same time. Custom plating is the longest-lead item on most BOMs.
  • Freeze colour at swatch approval. A colour changed after the dye lot is booked is a new dye lot.
  • Hold non-essential changes for the repeat. A stored specification makes the second order the right place for them.

The last line is the useful one. Most changes that arrive in round three are improvements rather than corrections, and improvements belong in the second order where they cost nothing extra and can be measured against real sell-through data rather than against a feeling.

What a Good Change Log Looks Like After Approval

The output of a sampling round is not only a sample. It is a change log that makes the next order cheaper, and it is the document an inspector refers back to when a dispute arises mid-run.

Change log structure for one approved round
FieldWhat it recordsWhy it matters later
R-numberSequential ID across all roundsLets any later query reference one change
Current and target valueTwo numbers per dimensional changeRemoves direction ambiguity
Round number and dateWhich round carried the changeTies the change to a sample
DecisionAccepted, rejected or deferred to repeatStops rejected changes reappearing
Effect on toolingNew die required, or noneExplains a cost line on the invoice
Effect on lead timeDays added, if anyExplains the schedule

Keep the log in one file and send it with the deposit. It is what converts sampling from a series of conversations into a record, and it is what allows a repeat order to skip sampling entirely where the specification is unchanged. The final gate is unchanged: inspection at AQL 2.5 under ISO 2859-1, which sets how many units are drawn from a lot of your size and how many defects of each class trigger a pass, a fail or a re-sort. A sourcing partner working to ISO 9001 will run the check against the approved sample and the change log together, and will attach the REACH Annex XVII, Proposition 65 and CPSIA documentation with the report. The full handbag private label process reference shows where this sits in the wider sequence.

What Comes Back: The Confirmation Table

Sending a clean list is half a round; the other half is what returns with the revised sample, and it is the part almost nobody specifies. Ask for a confirmation table: one row per R-number, with the measured result, the unit, and the photograph number that evidences it. A row that says done is not a confirmation, it is an intention. The table is what lets a buyer verify eleven changes in four minutes instead of holding the bag and guessing which of them were carried out.

Confirmation table for one revision round, and the entries that are not acceptable
FieldWhat it must containReject it when
R-numberMatches your list exactlyThe list has been renumbered
ResultA measured value with a unitDone, adjusted or improved
EvidenceA photograph or frame numberThere is no photograph
Not doneNamed, with the reasonThe row is simply blank

Set the clock on it as well. Acknowledgement within 24 hours, the confirmation table returned with the sample, and any blank row treated as a change that was missed rather than one that was unnecessary. Where a line genuinely cannot be executed, the reason belongs in the table on the day, so the next round starts from a shared record rather than from a memory of a phone call.

When Two of You Disagree About One Line

Most contradictory change notes are not carelessness; they are two people with different jobs looking at one sample. The fix is mechanical rather than diplomatic. Nominate one owner, give the internal disagreement a fixed budget of four working hours, and if it is not resolved inside that window the line moves to a deferred column with a D-number instead of blocking the round. A round waiting on an internal argument costs exactly as much as a round waiting on a material.

  • One owner, one channel. Two people emailing the same bench will eventually send opposite instructions.
  • A four-hour tie-break window. Unresolved after four hours, the line is deferred rather than held.
  • A deferred column with D-numbers. Deferred items are logged against the repeat order instead of being lost.
  • Never split the sample. One unit goes back with one list; a second unit with a second list produces two bags.

The deferred column is the part worth defending. Most round-three changes are improvements rather than corrections, and an improvement recorded as D04 can be measured against real sell-through data on the repeat, where it costs nothing extra. An improvement forced into round two costs a revision cycle and a decision made on a feeling.

Measure in the Same Conditions Every Time

Leather and coated materials move with humidity and temperature, and a dimension taken in a cold warehouse is not the dimension taken in a sample room. Condition the sample for 24 hours at 23 plus or minus 2 degrees and 50 plus or minus 5 per cent relative humidity before measuring, and use the same four tools every round: a steel rule, a calliper, a stitch gauge and a digital scale. A 34 cm tote can measure 6-8 mm wider in a damp room.

  • Condition first. 24 hours at 23 plus or minus 2 degrees and 50 plus or minus 5 per cent RH before any dimension is recorded.
  • Four tools, always the same ones. Steel rule, calliper, stitch gauge, and a scale reading to 1 g.
  • Measure empty and flat. A bag measured loaded or stuffed is a different bag from the one the line will build.
  • Record the room. Note temperature and humidity beside the measurement; a 6-8 mm swing is explainable, a 20 mm one is not.

Colour has the same problem. Evaluate under D65 or daylight at around 5000 K, never under warm shop lighting, and record a Delta E reading against the approved swatch rather than a description. A comment written as slightly dark will be interpreted differently by a bench working under 4000 K tubes, and the next round will contain a change nobody asked for.

Where a Round Actually Goes: The Hours Behind 3-6 Days

A consolidated round takes three to six working days, and knowing where those days go makes the number negotiable in the right places. Two to three hours go to reading and transcribing the list, twelve to sixteen hours to the bench work, three to four to checking each line against the sample, and three to four to packing and courier handover. The list itself took forty minutes to write. That ratio is the entire argument for consolidating before sending.

Where a revision round spends its time, and which steps can compress
StepHoursCan it compress?
Writing the revision list0.5-0.75It is already the cheapest step
Reading and transcribing the list2-3No, it scales with line count
Bench work12-16Slightly, with parallel stations
Checking each line against the sample3-4No, one pass per line
Packing and courier handover3-4No

The step worth attacking is the second one, and it is attacked by capping the list at twelve lines and writing each line so it can be transcribed without interpretation. Eleven changes in one list were executed in four working days; the same eleven arriving across four emails took thirteen, because that step ran four separate times and the bench restarted each time it did.

Send one list, get one clean revision

Send your revision list with photographs and dimensions. A private label coordinator will acknowledge within 24 hours, return the revised sample in 3-6 working days, and hold bulk at 35-50 days with inspection at AQL 2.5 under ISO 2859-1. Get a quote.

Frequently Asked Questions

How should I send design changes to a sample room?

As one numbered revision list, consolidated before sending, with one change per line and no more than 12 lines per round. Include annotated photographs, three angles per change, arrows numbered to match the list, and a marked-up dimension sheet. Send one channel with one named owner and never rely on a verbal instruction.

Why do follow-up emails cost so much time?

Because a bench cannot start a round it expects to be interrupted. Eleven changes sent across four emails over nine days produced six idle days, and the same eleven changes in one list were executed in four working days. Consolidating internally for 24-48 hours costs hours; sending piecemeal costs days.

What four fields does a change line need?

A sequential number such as R07, a location named from a datum, a current value and a target value, and a reference such as a Pantone code, a hardware code or a photograph number. Make the gusset deeper is an interpretation exercise; 100 mm to 110 mm at plus or minus 2 mm is an instruction.

How many changes can one revision round carry?

Twelve is a practical cap. Beyond that the round stops being a revision and becomes a redesign, and the probability of two lines contradicting each other rises sharply. If a list runs longer, split it: corrections now, improvements on the repeat order.

Which design changes are free and which reset the sampling clock?

Colour or hardware swaps within the approved range and branding moved within 5 mm are free. A dimensional change up to 10 mm on one panel is cheap. A material substance change is not a revision at all and resets sampling to 6-10 working days, and a closure change after tooling runs USD 150-600 for a new die.

How should sample photographs be annotated?

Three angles per change: a full view showing position, a mid view, and a close view with a steel rule or another object for scale. One arrow per change, numbered to match the revision list exactly. Shoot in daylight or at 5000 K, because a colour comment made under warm lighting will be evaluated under different light.

When is video better than a photograph?

For anything that moves. A strap that slips off the shoulder, a drawcord that will not feed, or a magnetic closure that pops under load are behaviours, and eight seconds of video shows them better than three sentences. Pair the clip with a written line so the record exists.

When should the specification be frozen?

Thirty days before the production slot begins, working backwards from the 35-50 day bulk window. That is roughly how long a hardware component with custom plating needs to be ordered and received. Freeze colour at swatch approval, because changing colour after the dye lot is booked means booking a new lot.

What happens if I change something after bulk starts?

It is priced per occurrence and carries schedule risk, because a change interrupts a line that is balanced across scheduled work. The cost is rarely the rework itself; it is the slot. Book the production slot when the deposit is paid rather than after approval, because slots tighten before Chinese New Year and during the autumn shipping rush.

How long should a revision round take?

Acknowledgement within 24 hours and a revised sample in 3-6 working days for a consolidated list. A full first sample takes 6-10 working days, rising to 12-15 for framed, woven or stacked constructions. Rounds slip when lists arrive piecemeal, not when the work is hard.

Why should I state what is already approved?

Because it stops a revision from quietly becoming a redesign. Without a statement of what is frozen, a bench told to change the gusset may also improve the pocket layout, and the next round contains a change nobody asked for and everybody has to pay for.

What document should a sampling round leave behind?

A change log with a sequential R-number, current and target values, the round number and date, the decision on each change, and the effect on tooling and lead time. Sent with the deposit, it is what lets a repeat order skip sampling entirely where the specification is unchanged.

People Also Ask

Can two people on my team send changes separately?

No. Two people emailing the same sample room will eventually send opposite instructions, and the bench cannot adjudicate between them. Nominate one owner and consolidate internally before sending.

Is a voice message acceptable for a change?

No. Anything not written down will eventually be missed, and the miss is always discovered at the worst moment. Record it in the numbered list even if a call happened first.

Should I hold improvements for the second order?

Usually yes. Most round-three changes are improvements rather than corrections, and improvements cost nothing extra on a repeat where the specification is stored, and can be judged against real sell-through data.

What is the fastest way to lose a season?

Changing the material substance in round three. It resets sampling to 6-10 working days and moves drape, hand feel and yield together, none of which can be predicted from the previous sample.

Sources and further reading

Every figure on this page is either our own programme data or attributed below. Open the originals before you write them into a specification.