Bag Print Quality: Colour Fastness, Cracking and Wash Testing Explained

Most creators approve a printed sample by looking at it under good light and nobody ever teaches them what to ask for instead. That is understandable, because printed fabric either looks right or it does not, and the numbers behind it feel like laboratory trivia until a backer posts a photograph of a white shirt stained blue. Our production team has reviewed enough returns to know the pattern: colour complaints almost always trace back to a missing numerical target, and cracking complaints trace back to choosing a print method whose elasticity did not match the fabric. This guide covers both, with the specific test codes and acceptance values worth writing into a tech pack.

What Colour Fastness Actually Measures

Colour fastness describes how well a coloured surface resists change when it is exposed to something, and the crucial detail is that there is no single number. It is always a resistance to a specific agent, and each agent has its own test method, its own scale and its own realistic ceiling. A pigment that scores brilliantly against washing can score badly against light, and neither result predicts the other. This is why the question what is the fastness is unanswerable, and the question what is the rub fastness rating is the one worth asking.

The two families you will meet are the ISO 105 series and the AATCC methods, and they are broadly comparable but not interchangeable in the numbers they produce. Most mills we work with report ISO results as the default. Within ISO 105 there are two assessment scales that people mix up constantly: the grey scale for colour change, designated ISO 105-A02, grades how much the specimen itself faded, while the grey scale for staining, ISO 105-A03, grades how much colour transferred onto an adjacent undyed cloth. Those are different questions with different answers on the same test.

Understanding that split resolves most arguments. A printed panel can score 4 for colour change, meaning it barely faded, and 2 for staining, meaning it deposited plenty of colour onto the adjacent fabric, and that combination is exactly how you end up with a reward bag that looks fine itself and ruins whatever it rubs against. Always ask for both figures. Our pillar overview of custom bags for Kickstarter campaigns explains where these specifications sit in the wider sampling sequence, and how long each stage realistically takes.

• Colour change grades whether the specimen itself faded, assessed to ISO 105-A02.
• Staining grades colour transferred to adjacent fabric, assessed to ISO 105-A03.
• Agent specificity means wash results say nothing about light results and vice versa.
• Ask for both numbers on every fastness test, because one alone is ambiguous.

Reading the 1 to 5 Grey Scale Without Guessing

Most textile fastness results are reported on a five-point scale, where 5 means no perceptible change and 1 means severe change. It is not linear and it does not behave like a school grade, which is the single biggest source of confusion. The practical interpretation runs roughly like this: 5 is no change, 4 to 5 is a change noticeable only to a trained eye, 4 is slight, 3 to 4 is noticeable but usually acceptable, 3 is clearly visible, and anything at 2 or below is a complaint waiting to happen.

Half steps matter and they are not rounding errors. 4 versus 3 to 4 is the difference between negligible and noticeable, which is why most commercial specifications quote values like 4 rather than whole preferences. The interesting economic point is that each step up the scale is progressively harder and more expensive to achieve, so specifying 5 across the board will raise your cost for a benefit nobody perceives, while specifying 3 will generate complaints you did not budget for. The sensible operating band for reward bags sits between 3 to 4 and 4 to 5 depending on the agent.

There is one exception to the five-point convention worth flagging: light fastness uses the blue wool scale, which runs from 1 to 8 and behaves quite differently. On that scale a printed panel rated 4 is genuinely poor for outdoor use and 6 is the realistic commercial target for a product that will see regular daylight. Getting these two scales confused produces specifications asking for light fastness 5, which sounds demanding and is in fact a mediocre requirement. We return to this in the light fastness section, and the equivalent numbers for other surfaces are covered in print and embroidery QC including registration and shrinkage.

Rub Fastness: The Test Protecting Your Backer's Clothing

If you only fund one test, fund this one, because it predicts the failure mode that generates the angriest messages. Rub fastness, also called crock fastness, is measured to ISO 105-X12 by rubbing a standard cotton cloth across the printed surface under a defined load for a defined number of cycles, and assessing staining of that cloth on the grey scale. The method covers both dry rubbing and wet rubbing, and the two results should always be reported separately, because they routinely differ by a full point or more on the same print.

Wet results being lower is normal and not necessarily a defect. Moisture swells fibres and mobilises surface colour, so a wet rating one point below dry is common across most print technologies. What matters is the absolute floor. For reward bags we generally target dry 4 or better and wet 3 to 4 or better on printed outer panels, and we push toward dry 4 to 5 on anything that sits against clothing for hours, such as the back panel of a backpack or the underside of a strap. Anything arriving at wet 3 or below needs a conversation before approval.

Several fixable variables move the result. Incomplete curing is the most common cause of a poor rub score and has nothing to do with the ink chosen. Excess surface deposit from too-heavy ink coverage rubs off in the first week and then mysteriously improves, which is why a factory can genuinely pass a test on a washed sample while the shipped product fails. Softener and finishing residues also interfere, particularly on dark, heavily printed panels. The cheapest control is to require testing after the final wash rather than before it, and our companion guide to reducing bleed and cracking on printed fabric covers the practical fixes in production order.

Wash Fastness and Honest Care Instructions

Wash fastness is measured predominantly to ISO 105-C06, which launders the specimen with a standard detergent and a set of adjacent multi-fibre fabrics, then grades both colour change and staining. The method has numbered variants specifying temperature and liquor conditions, and C06 at 40 degrees Celsius is the common commercial default, with higher temperature variants available for products marketed as machine washable in hot cycles. The number is meaningless without the variant, so always ask which one was run.

Here is the part most creators get wrong, and it costs them the most goodwill: you do not have to make the bag machine washable, you have to tell the truth about it. A structured backpack with foam, interfacing and a coated shell is not a machine-washable object, and the honest care instruction is surface clean with a damp cloth. What you must not do is sell a bag with a 40 degree wash fastness rating of 3 and then write machine washable on the campaign page, because the failure will be yours and not the print's.

When the bag genuinely is washable by design, usually a tote, a pouch or a simple drawstring sack, target colour change 4 or better and staining 4 or better at the temperature you intend to print on the care label. Then test the care label claim, not just the fabric, because shrinkage of the surrounding material frequently distorts a print that itself passed every colour test. Wash testing typically needs 3 to 5 working days for a third-party report, which fits comfortably inside the standard 6 to 10 working day sampling window if you request it at the same time rather than after approving the sample.

Light Fastness: Sunlight, Shop Windows and Six Months of Use

Light fastness uses the blue wool scale from 1 to 8 under ISO 105-B02, where eight reference dyed wool standards are exposed alongside the specimen and the rating is set by how many have faded comparably. This is why the scale is not five points, and why quoting light fastness as if it were on the same scale as rub fastness produces specifications that sound strict and are weak. For a bag that lives indoors and sees occasional daylight, 4 to 5 is a reasonable commercial floor.

The exposure case matters more than the number. A commuter backpack spends its life in indirect daylight, so its realistic exposure is modest. A beach bag, a cycling bag or anything designed to sit on a sunny balcony is in a completely different regime, and those products need 6 or better, which usually means choosing different pigment chemistry rather than asking the same ink to perform better. Very occasionally creators discover this the expensive way, with a warranty claim arriving after the second summer.

There is also a retail reason nobody thinks about until later. If your plan post-campaign is to wholesale into shops, your bags will sit in a window under continuous light for weeks, and fabrics that pass a casual use case can visibly shift in that environment. Retail exposure is materially more aggressive than domestic exposure. Because certain reds, oranges and fluorescent brights are structurally weaker on light fastness than earth tones and blacks, choosing palette colours with this in mind early is far cheaper than reformulating later, and the method comparison in custom printed bag fabrics including sublimation and screen indicates which technologies hold brights best.

1. Light fastness runs on a blue wool scale of 1 to 8, not the 1 to 5 grey scale used elsewhere.
2. Target 4 to 5 for indoor and commuter use, and 6 or better for sustained outdoor exposure.
3. Test retail window exposure separately if wholesale is part of your post-campaign plan.
4. Expect bright reds, oranges and fluorescents to be the weakest colours in any palette.

Why Prints Crack: Elongation, Film Thickness and Flex

Cracking is not a colour problem and it will never show up in a colour fastness report, which is precisely why it surprises people. A print cracks because the ink film sitting on top of the fabric cannot stretch as much as the fabric underneath it can. When a bag panel bends, folds or gets crushed into a suitcase, the substrate elongates, and if the ink layer has lower elongation than the textile it fractures. Every visible crack is a mechanical mismatch, not a colour error.

Three factors control it and they compound. The first is deposit thickness, because thicker films crack sooner, so heavy ink coverage is a risk rather than a quality signal. The second is fabric stretch, because a four-way stretch knit and a coated woven behave nothing alike under the same fold, and the same artwork behaves differently on each. The third is where the print sits, since a panel over a foam pad or a stiffener flexes far less than one over an unsupported area, while a print crossing a fold line or a binding edge has essentially no chance.

The standard screening check is simple and worth requesting by name: flex and fold the printed panel sharply at least twenty times, then inspect under magnification for hairline crazing. For the more formal route, adhesion is assessed by cross-hatch tape testing to ASTM D3359, which grades how much of a cut grid lifts, and durability under rubbing is assessed by Martindale cycles on the printed face. We generally look for no visible crazing through a magnifier after twenty folds and no more than modest grid loss at the ASTM D3359 4B level or better. Our comparison of screen printing versus heat transfer for reward bags shows which methods carry lower cracking risk on which substrates.

The Test Protocol to Request Before Bulk Approval

Testing only works if it is requested before the sample is approved, not after. Once gold seals are signed off, any retest becomes a negotiation about cost and schedule rather than a quality control, and it lands in the same window as your campaign shipping promise. The efficient pattern is to attach the full test list to the sample approval form as a single block, so that a sample cannot be approved without results attached, and to name both the standard and the minimum grade in writing.

There is also a sequencing point that saves real time. Third-party laboratory reports generally take around 3 to 5 working days, while the internal production floor checks for rub, adhesion and flex take hours. Run the internal checks first on every colourway, then send only the combinations closest to the limit to the external laboratory, which typically halves both the cost and the elapsed time. Given that sampling runs 6 to 10 working days for ordinary builds and 12 to 15 working days for complex ones, requesting tests at the start rather than the end is the difference between a comfortable approval and a rushed one.

Finally, test production-intent samples, not presentation samples. It is a normal human behaviour to submit the best-made article from the bench, and it is also completely unrepresentative. Ask specifically for a sample produced on the same line and with the same settings intended for bulk, including the same curing parameters, because curing is exactly the variable that moves rub fastness and is most often adjusted between a presentation piece and a production run. Then re-verify at bulk with AQL 2.5 to ISO 2859-1 inspection and a retained reference from the approved sample to compare against.

Setting Numerical Acceptance Criteria in the Tech Pack

A print specification that says colour must match the approved sample and must not fade is unenforceable, because every party will interpret it differently and none can test it. Replace qualitative language with numbers and you convert an argument into a checklist. The minimum set for a printed reward bag fits comfortably into six lines: print method, colour reference with tolerance, rub fastness dry and wet, wash fastness with variant and temperature, light fastness, and adhesion or flex criterion.

Add two lines that are almost always missing and cause almost all remaining disputes. The first is which ASTM D2244 colour tolerance applies, together with a numerical delta E limit, because without it any visual difference becomes a matter of opinion. The second is critical placement, specifically whether a print crosses a fold, binding or seam, since the standard fastness targets assume the print sits on a flat supported panel. A print that crosses a fold needs different expectations and usually different artwork geometry.

Then decide what happens on failure, before it happens. Write into the document whether a failing result stops shipment, triggers a re-cure attempt, or permits shipment with a discount, and who pays for retesting. Sorting that out proactively takes fifteen minutes and removes entirely the worst category of dispute, which is the one conducted while a container is already booked. The artwork side of this is equally specifiable, and our guide to artwork files for AI, EPS, PDF and PNG output sets out the file standards that keep a clean print achievable in the first place.

Troubleshooting the Five Most Common Print Failures

The first complaint is colour bleeding onto clothing, which is almost always a rub fastness value below roughly 3 to 4, and almost always traces to incomplete curing rather than the ink selected. The second is cracking at folds, which we have covered: reduce deposit thickness, switch to a method with greater elongation, or move the artwork away from the fold. The third is fading unevenly, which is normally light fastness below 4 on the blue wool scale attacking one specific pigment in the artwork, typically a red or a bright.

The fourth is print washing off the panel, meaning an adhesion failure. This usually points to either insufficient surface preparation on a coated or water-repellent fabric, or a cure temperature that never actually reached the target even though the machine reported it. The fifth, and the one that confuses people most, is colour differing from the approved sample when every individual test passes, because comparing across dye lots and print runs without a retained physical reference is comparing memories. Hold a signed standard and compare physically under consistent lighting.

All five share one characteristic: they were visible in testing and nobody was funded to look. Fix that by putting at most USD 150 to 400 of testing into almost any reward bag programme, a sum that is trivial beside the cost of a reprint and irrelevant beside the cost of a bad reviews cycle. Our team schedules those checks inside the normal 6 to 10 working day sample window, with bulk following 35 to 50 days after approval, so a properly tested programme costs days rather than weeks.

Frequently Asked Questions

People Also Ask