What Shade Band Should a Reorder Colour Match Sit In?
A colour reference names a target; it does not set a tolerance. What keeps a reorder predictable is a named physical reference, an instrumental band agreed before production, and a gray scale fallback when two people read the same panel differently. This article covers all three, plus how the band is reviewed once a programme has several runs behind it.
['MOQ 500 pcs per style and colourway; 300 pcs open to discussion on a repeat.', 'Sampling 6-10 working days from a signed-off tech pack.', 'Bulk production 35-50 days from written sample approval.', 'Inspection at AQL 2.5, ISO 2859-1, General Inspection Level II.', 'T/T 30/70, FOB Xiamen; CIF and DDP available.', 'A colour reference names a target; a band sets the tolerance.', 'Instrument reading for the arithmetic, gray scale for the argument.']
Colour is the property buyers care about most and specify least precisely. A style is described by a name or a code, the sample is approved, and the reorder is booked against that name. Two years later the two sides are describing the same colour in different words and neither can produce a number.
The fix is not a better name. It is a reference plus a band: a physical panel both sides hold, a number that says how far the reading may move from it, and a method that says how the reading is taken. Those three convert colour from a conversation into a measurement.
None of that requires a laboratory on site. What it requires is one decision made early and written down: which reference, which band, which method, and who reads it. Four answers, agreed before the season is booked, and colour stops being the property that a reorder is argued about.
What Does a Colour Reference Actually Specify?
A reference system gives a name to a colour so two people can talk about it. It is invaluable for naming a target and it does not, by itself, say how far a production lot may differ from that target. That distance is the band, and it has to be agreed separately.
It also does not specify the substrate. The same reference name on polyester, on nylon and on a coated fabric produces three different colours, because the dye uptake, the surface and the finish differ. A reference is a target for a material, never a material in itself.
There is a third limitation worth naming. A reference system is revised and re-issued, and a name from an older edition does not always sit where the current edition puts it. Record which edition a reference was taken from, and keep the physical panel, because the panel is never re-issued.
The same caution applies to a reference quoted from a screen. A value read off a display depends on that display, its calibration and its age, which is why a digital reference is a development convenience and never the thing a dispute is settled against.
Where a buyer works across several programmes, agree one naming convention and one illuminant set for the whole range rather than per style. A house rule removes an entire category of argument, and it is the sort of decision that is much easier to make once than five times under time pressure.
Why Is a Printed or Digital Reference Not a Standard?
Because a printed swatch and a screen are both rendered, and a rendering depends on the ink, the paper, the profile and the display. Two people can hold the same printed card and be looking at a slightly different colour, which is precisely what a standard is supposed to prevent.
| Reference type | What it fixes | What it cannot do | Where it belongs |
|---|---|---|---|
| Named system reference | A shared name for the target colour | Set a tolerance or specify the substrate | On the spec file as the naming convention |
| Retained fabric panel | The actual approved colour on the actual substrate | Be reproduced from a description | With the gold sample, on both sides |
| Instrument reading | A number under a named illuminant | Describe what a customer sees in a shop | On the comparison sheet, per run |
| Gray scale grade | A graded judgement of perceived difference | Replace an instrument when the band is tight | As the fallback when two people disagree |
| Printed card | A quick shared visual cue | Serve as the standard, or survive a screen change | In the meeting, not on the file |
| Digital swatch | A fast way to shortlist shades | Carry a tolerance or a substrate | During development only |
How Is an Instrumental Band Written?
As a colour difference number against the retained panel, with the illuminant and the observer angle named, the instrument named, and the number of readings per unit stated. A band written without those four is a band that will be re-litigated the first time it is used.
There is no published commercial limit for bags. The band is an agreement between the two parties, and the practical way to set the first one is to read ten units from the approved run, take the spread, and set the band just outside it. Tighten it later, once two consecutive runs sit comfortably inside.
Write the outcome on the same line as the band. Sorting, rework, acceptance on one run with a written exception, or rejection of the lot are four different answers, and deciding which applies after the cartons have landed is how a small colour difference becomes a large commercial one. One extra column prevents that outcome.
Say who reads and when. The reading that settles a dispute is the one taken at inspection, before the cartons are sealed, by a named grader, against the named panel. A reading taken later, at the destination, by someone else under a different lamp, will not match the first one and should not be expected to.
What Role Does the Gray Scale Play?
The gray scale grades a perceived difference in steps rather than producing a number. It is slower and more subjective than an instrument, and it is closer to what a customer experiences, because a customer is a person and not a spectrophotometer.
Use it as the fallback. Where the two sides read the same panel differently, a gray scale grade gives a graded answer that both can accept, and it is quicker to agree on than a dispute about which instrument setting was correct.
- Reference
- The retained panel from the approved run, held by both sides
- Number
- A colour difference limit under a named illuminant
- Observer angle
- Named, so two instruments produce comparable numbers
- Readings
- How many readings per unit, and at which named position
- Fallback
- A gray scale grade, and who grades when the two sides differ
- Illuminants
- Which light sources the approval was read under
- Outcome
- What happens when a reading sits just outside the band
How Many Units and How Many Readings?
Units follow the sampling plan for the lot. Under ISO 2859-1 at General Inspection Level II and AQL 2.5, a lot of 501 to 1,200 units takes a sample of 80 and a lot of 1,201 to 3,200 takes 125. Read each unit at the same named position, and take at least three readings per unit at that position.
Take three readings per unit because a panel is not uniform. A single reading can sit at the edge of a shade variation within one panel, and three readings give a mean and a spread. The spread is the useful part: a unit whose three readings vary widely is a unit from a problem lot even when the mean passes.
How Do You Handle a Multi-Panel Colourway?
Give each panel its own reference and its own band. A colourway with a body fabric, a contrast panel and a binding has three colour decisions, and comparing them all against one reference hides the one that moved.
Then retain an assembled reference as well. Two panels can each match their own reference and still look wrong together, because the relationship between them is part of what was approved. A retained assembled swatch is the only thing that captures that.
What Happens With Trim and Hardware Colour?
Trim is harder than fabric because it is bought rather than dyed to order. A webbing colour, a zipper tape and a plated D-ring each come from a separate supply chain with its own tolerances, and none of them was necessarily matched to the body fabric in the first place.
Register each trim colour to the same discipline: a retained physical sample, a band, and a named light source for comparison. Where a trim item cannot be matched closely, say so on the file and agree the accepted difference in writing rather than leaving it to be discovered at goods-in.
| Colour decision | Who controls it | Reference held | Band written |
|---|---|---|---|
| Body fabric | Dyehouse, matched to the retained panel | Cut panel from the approved run | Instrumental, with a gray scale fallback |
| Contrast panel | Dyehouse, matched separately | Cut panel, labelled by position | Instrumental, separate band |
| Binding and tape | Trim supplier, by code | Retained length of the approved tape | Gray scale against the retained sample |
| Zipper tape | Fastener maker, by code | Retained zipper from the approved run | Gray scale, plus the code on the register |
| Plated hardware | Plating bath, by finish code | Retained slider and ring | Visual under a named light, plus the code |
| Print and logo | Print method and ink | Retained printed panel | Position and shade, both recorded |
How Is the Band Reviewed After Several Runs?
Once a year, on the fifth-reorder test or at the annual review, read the trend of the recorded numbers rather than the pass and fail lines. Where every run sits well inside the band, tighten it. Where runs cluster at the edge, either widen it honestly or fix the process that is producing the cluster.
A band that every run passes comfortably is too wide to be doing any work, and a band that half the runs fail is not a control. Both are visible only on a chart with several runs on it, which is why the readings have to be recorded rather than only the result.
Review the reference at the same time. A band can be correct and the panel it points at can be wrong, because the panel has aged. Compare the working panel against the archive panel once a year and replace the working one where the two differ, before rather than after the season is booked.
Finally, review the illuminant set. Where a programme has entered a market with different retail lighting, a band agreed under one source may not describe what the customer sees. Adding a second or third illuminant at review time is far less disruptive than discovering the need for one after a shipment has been rejected.
What Should the Buyer Send Alongside the Reference?
Four things: the named reference, a physical panel if one exists, the substrate specification, and the light source the colour will be viewed under in the market. The last one is the most often omitted and the most useful, because a shade matched under the wrong light is a shade that will be rejected in the shop.
Four things to send with a colour
- The named system reference, as the naming convention for the target.
- A physical panel or swatch from anything already approved.
- The substrate: fibre, weave, coating and finish, named on the register.
- The light source the product will be viewed under where it is sold.
What Goes Into the Colour Record?
One sheet per run per colour decision: the reference used, the illuminants, the instrument, the readings with the mean and the spread, the gray scale grade where one was taken, and the pass or fail line. Seven fields, and the readings are the part that makes it a record rather than a result.
Add the dye-lot number to the same sheet. Colour and dye lot are the same story told two ways, and a record that carries both lets a reading be traced back to the bath that produced it rather than leaving the two in separate folders that never meet.
Keep the sheet for the life of the style. A colour record covering six runs is the fastest answer to the question that eventually arrives in every programme - whether the colour has moved - and it is also the only document that can distinguish one bad lot from a slow drift across several.
A reference names a target and a band sets the distance. Write both down, and colour stops being the thing a reorder is argued about.
Send us the reference and the band
Send the named reference, the panel you hold and the light the bag will be sold under. We coordinate production with a lab dip matched to the panel per lot, bulk in 35-50 days from written sample approval.
Frequently Asked Questions
What is a shade band?
A colour difference limit agreed in writing before production, expressed as an instrumental difference from the retained panel under a named illuminant and observer angle, with a gray scale grade as the fallback. There is no published commercial limit for bags; the band is an agreement.
Why is a printed colour reference not enough?
Because a printed swatch and a screen are both rendered, and the rendering depends on the ink, the paper, the profile and the display. Two people can hold the same card and see a slightly different colour, which is what a standard is supposed to prevent.
How do you set the band for the first time?
Read ten units from the approved run, take the spread, and set the band just outside it. Tighten it later, once two consecutive runs sit comfortably inside. A band the process cannot hold becomes a permanent argument rather than a control.
What role does the gray scale play?
It grades a perceived difference in steps rather than producing a number. It is subjective between graders and closer to what a customer sees, so use it as the fallback when the two sides read the same panel differently.
How many readings should be taken per unit?
At least three, at the same named position, recording the mean and the spread. A panel is not uniform, and a unit whose three readings vary widely comes from a problem lot even when the mean passes.
How many units are checked for colour?
The sample follows the standard plan for the lot size. Under ISO 2859-1 at General Inspection Level II and AQL 2.5, a lot of 501 to 1,200 units takes a sample of 80, and a lot of 1,201 to 3,200 takes 125.
How is a multi-panel colourway handled?
Give each panel its own reference and its own band, then retain an assembled swatch as well. Two panels can each match their own reference and still look wrong together, because the relationship between them is part of what was approved.
Why is trim colour harder than fabric colour?
Because trim is bought rather than dyed to order. A webbing colour, a zipper tape and a plated ring come from separate supply chains with their own tolerances, and none was necessarily matched to the body fabric in the first place.
What should a buyer send with a colour?
Four things: the named system reference, a physical panel from anything already approved, the substrate specification including fibre, weave, coating and finish, and the light source the product will be viewed under in the market.
How often should the band be reviewed?
Once a year, at the annual review or the fifth-reorder test. Where every run sits well inside, tighten it; where runs cluster at the edge, widen it honestly or fix the process producing the cluster.
What goes into a colour record?
One sheet per run per colour decision: the reference, the illuminants, the instrument, the readings with mean and spread, the gray scale grade where taken, and the pass or fail line. The readings make it a record rather than a result.
Can the same band apply to every colourway?
The same method can, but not the same number. Dark shades, bright shades and metallics each behave differently under instrument reading and under the eye, so each colourway needs its own band agreed against its own retained panel.
People Also Ask
What is an acceptable colour tolerance for reorders?
A band agreed in writing before production: an instrumental colour difference from the retained panel under a named illuminant and observer angle, with a gray scale grade as the fallback. Set the first one from the spread of ten approved units.
How do you match Pantone colours on fabric?
Use the named reference to communicate the target, then match physically: dye a lab dip, compare it to the retained panel under the named illuminants, adjust, and only then run the bulk lot. The substrate changes the result, so never match by name alone.
Why does the same Pantone look different on different materials?
Dye uptake, surface texture, coating and finish all change the apparent colour of the same target. A reference names a colour for one material; it does not transfer across substrates without a physical match.
What is a gray scale used for in colour matching?
Grading a perceived colour difference in steps. It is subjective between graders but close to what a customer sees, so it works best as the fallback when an instrument reading is disputed.
How many colour readings should I take per sample?
At least three at the same named position, recording mean and spread. A single reading can sit at the edge of a variation within the panel; the spread tells you whether the lot itself is uniform.
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.
- Pantone Matching System — colour standards and referencesPantone
- CIE 15 colorimetry and the CIEDE2000 colour-difference formulaInternational Commission on Illumination
- AATCC evaluation procedure 1 — gray scale for colour changeAmerican Association of Textile Chemists and Colorists