Production

Reducing misprints: a QC checklist for apparel decorators

Most misprints are caught by five checks: count and inspect blanks before decorating, verify the approved proof against the setup, sign off a first-off print before running the batch, wash-test the cure on a sample, and count again at pack. Put the checks before the expensive step rather than after it, and record who signed each one.

Most reprints trace back to five checks that take under two minutes. The checklist, and where in the run to put each one.

Two workers in a textile factory sorting fabric at a table under bright lighting.
Photo: EqualStock IN / Pexels

Reprints are the most expensive thing a print shop does. You pay for the blanks twice, the labour twice, and often the freight twice, on a job that is now late. And almost all of them are caught by checks that cost under two minutes.

The principle: check before the expensive step, not after it.

The five checks

1. Blanks in: count and inspect

Before anything is decorated. Count against the purchase order, check sizes and colours, look for mill faults, dye variation between cartons and mislabelled sizes. Finding a short shipment now costs a phone call. Finding it after decorating 46 of 50 costs the job.

2. Setup against the approved proof

Position, size, garment colour, ink colours, quantity per size. Read them off the approved proof, not off memory or a verbal. This is where "we printed the old version" is prevented, and it is why proof versioning matters.

3. First-off sign-off

Print one. Stop. Check it against the proof for registration, opacity, position and placement height. Then run the batch. A shop that skips this is betting a whole run on a setup nobody has looked at.

4. Cure verification

Stretch test on the first-off, and a wash test whenever the ink, garment or dryer settings are new. Under-cure is invisible on the day and obvious after the customer's first wash, which is the worst possible time to find out.

5. Count and check at pack

Count by size against the order. Look at every piece as it folds, front and back. Set aside seconds rather than hoping nobody notices. Pack, label, and record the count.

Where each check sits

Check Position in the run Cost if skipped
Blanks in Before decoration Whole batch of wrong garments
Setup vs proof Before first print Whole run, wrong version
First-off After one piece One garment vs the batch
Cure test After first-off Customer's washing machine
Pack count Before dispatch Freight twice, credibility once

Make the checks recorded, not remembered

A checklist that lives in someone's head is a preference. Two things turn it into a process:

The seconds bin

Have one, use it, and price accordingly. A shop with no seconds is either extremely good or shipping its mistakes. Building your real reject rate into your cost per unit means the occasional spoiled garment is already paid for rather than eating a job's margin.

Count the cost of a reprint properly

A reprint is not one blank. It is the blank, the ink, the press time, the setup if screens came down, the delivery you now have to expedite, the admin of explaining it, and the slot it takes from a job that was already scheduled. Costed honestly, a single reprinted 24-piece job can wipe out the margin on two similar jobs.

Cost line Typical on a 24-piece rerun
Blanks written off £58
Ink and consumables £9
Press and labour, setup plus run £70
Expedited freight £18
Admin and customer handling £15
Displaced work The real one, hardest to price

That is why the checks above belong before the expensive step. Two minutes at first-off is the cheapest insurance in the building.

Make the record automatic rather than diligent

Checklists on paper work for a fortnight and then live in a drawer. What survives is a check that is part of moving the job forward: a stage the job cannot leave until someone has advanced it, with a name and a timestamp attached.

In practice that means QC is a stage on the board rather than a habit, the approved proof is attached to the order so the setup check is against the exact version the customer signed, and advancing the job is a scan or a tap that records who did it. If staff also clock on and off the job, you get the other half of the picture: which jobs consumed rework hours, visible in the job costing report as jobs whose labour ran far past the norm for their size.

That pairing, a recorded check plus recorded hours, is what turns quality from an argument into a number.

Track five cause codes, review monthly

  1. Artwork. Wrong version, missing bleed, low resolution, unapproved change.
  2. Setup. Registration, placement, wrong garment or size pulled, wrong ink.
  3. Material. Faulty blank, dye migration, inconsistent fabric.
  4. Cure. Under-cured, over-cured, dryer drift.
  5. Handling. Marked, folded wet, contaminated, damaged in pack.

A month of codes will usually show one dominant cause, and it is often material or artwork rather than the press work everyone assumes. Material faults point at a blanks decision; artwork faults point at the approval process, where the fix is a clear approved version rather than a longer checklist.

The weekly number to watch

One line on your weekly review: rejects as a percentage of garments decorated, and the reprint cost in money. Both trending down is a shop getting better. Reject rate flat while volume rises is a shop that will hit a wall in rush season. See print shop KPIs for where this sits among the other numbers worth reading weekly.

Printer's Friend attaches the approved proof to the order so the setup check is against the exact version the customer signed, and QC is its own stage on the floor board with a scan-in so the count is recorded rather than recalled. For where QC sits in the whole path, see the nine stages.

Free download

Get the pricing pack

One formula per decoration method, a worked example, and the seven numbers to watch. Free PDF.

  • One formula per decoration method
  • 100 black tees, priced line by line
  • The seven numbers to watch weekly

Questions

What is a first-off and why does it matter?
The first finished piece of a run, checked against the approved proof for position, size, colour and cure before the rest run. It converts a whole-batch error into a single-garment error, which is the highest-value two minutes in the shop.
How do I test the cure without ruining a garment?
Stretch the print firmly; a properly cured plastisol print stretches with the fabric and does not crack. For confidence on a new ink or garment, wash-test a sample piece several times before shipping the order.
Should I charge the customer for their own artwork mistakes?
If they approved a timestamped proof that shows the error, the reprint is chargeable, and you should say so calmly and offer a goodwill discount. That only works when the approval is recorded against a specific version.
What reject rate should I expect?
Between 1 and 3 percent of garments decorated is normal for a well-run shop across mixed methods, higher for DTG on darks and for anything with names. The absolute number matters less than whether you know it: a shop that cannot state its reject rate is absorbing an unknown cost on every quote.
Who should sign the first-off?
Whoever is accountable for the run, and it should be recorded with a name and a time rather than a nod. On a busy floor the useful pattern is that the operator prints it and a second person compares it against the approved proof, because the person who set the job up is the least likely to spot their own mistake.
How do I stop the same fault recurring?
Record a cause code on every reject, not just a count. Five codes are enough: artwork, setup, material, cure, handling. A month of coded rejects tells you which single change would remove the most of them, and it is rarely the one people argue about.

Run the shop, not the chaos.

Printer's Friend turns every job in this article into a tracked order: quote, artwork approval, production stage, invoice.

Start free

Read next

Print shop workflow: quote to dispatch in nine stages → Shop floor layout for a two-press operation → How to cut artwork approval time from days to hours →