How to Handle a Production Problem or Defective Bulk Order

You approved the sample.

You paid the deposit.

You waited eight weeks.

And now the boxes are open in your storage unit, and roughly a third of the run has crooked hems, off-tone dye, or print that cracked on the first stretch.

Before you fire off an angry voice note at 2 am, stop.

What you do in the next 48 hours decides whether you recover most of your money or eat the loss.

This guide walks through the five decisions in order.

Here’s the uncomfortable truth most founder blogs skip: a defective bulk order is not a rare disaster.

On a first production run with a new supplier, some level of defect is closer to the norm than the exception.

The founders who lost money aren’t the ones who got defects; they’re the ones who reacted emotionally, destroyed their own leverage, and negotiated from a position of weakness they created themselves.

This is written from inside the manufacturing side.

On the factory floor, we see how these disputes actually resolve, which claims get a full remake and which get ignored, and why.

The difference almost always comes down to the same five decisions below.

Work through them in order, and you’ll hold your margin.

Skip to “demanding a refund,” and you’ll usually walk away with nothing.

01:

Document the problem with evidence:

Your first instinct will be to message the supplier:

“The order is bad, I’m really disappointed, and this is unacceptable.”

Resist it. That message does nothing for you and quietly signals that you’re operating on emotion, not evidence.

In a dispute, evidence is leverage, and evidence degrades fast.

Before you contact anyone, treat the shipment like a claim you might have to prove to a stranger.

Because you might go to the supplier’s owner, to a payment platform’s dispute team, to a third-party inspector, or eventually to yourself when you’re deciding whether you have a real case or just bad luck.

Freeze the goods exactly as they arrived:

Do not wash, alter, re-tag, sell, or ship a single defective unit until the claim is settled.

The moment you modify the goods, the supplier can argue the damage was yours.

Keep the original packaging, poly bags, cartons, and shipping labels.

If units arrived damaged in transit rather than defective from production, that packaging is the difference between a factory claim and a freight claim, two completely different conversations.

Photograph everything, the right way:

Bad photos sink good claims.

Shoot in daylight or neutral light, and capture three layers of evidence for each defect type:

• The wide shot: the defect in context on the whole garment, so it’s clearly your product.
  

  ---

• The close-up of the defect itself, sharp and filling the frame (a broken stitch, a print crack, a stain, a misaligned panel).
  

  ---

• The measurement showed the defect against a tape measure or ruler, so a sizing or placement error is undeniable in numbers, not adjectives.

Then photograph the same point on a good unit for contrast.

“Some are bad” is an opinion.

“The print sits 3 cm lower than spec on 40 of 150 units; see attached” is a claim a factory has to answer.

Count and categorize: Don’t estimate

“A lot of them are defective” is worthless in a negotiation.

A defect rate is everything.

Pull a representative sample (or inspect the full run if it’s small enough) and sort every defect into three buckets the whole industry already recognizes:

• Critical: unsellable and, in some categories, unsafe (broken zipper on outerwear, a needle fragment, or a seam that fails on first wear).
  

  ---

• Major: a reasonable customer would return it (visible stain, wrong color, holes, broken stitching, sizing off-spec).
  

  ---

• Minor: a small flaw a customer likely wouldn’t reject (a loose thread, a faint imperfection).

This isn’t arbitrary. It maps to AQL, the acceptable quality limit.

The acceptance-sampling standard the apparel industry runs on, defined in ISO 2859-1 (and its US twin ANSI/ASQ Z1.4).

A typical apparel inspection is set at AQL 2.5 for major defects and 4.0 for minor ones.

In plain terms: a certain small number of minor flaws are tolerated by the standard; go above the limit, and the lot officially “fails.”

When you frame your claim in AQL terms, you stop sounding like an upset customer and start sounding like a buyer who knows the rules the factory already agreed to play by.

Compare against the contract: Not your memory

Here’s where most founders quietly lose: they claim the goods are “wrong” without a written standard to measure against.

The contract in apparel is not your DMs; it’s your tech pack, the graded measurement/spec sheet, and the signed-off pre-production sample.

If your tech pack specified a 320 GSM fleece and you got 240, that’s a documented breach.

If you only “assumed” heavier fabric, you have nothing.

Lay the defective units next to the approved reference and photograph the difference.

A claim anchored to an approved spec is nearly unarguable; a claim anchored to your expectations is nearly unwinnable.

Timestamp and raise it fast:

Record the PO number, batch/lot number, date received, and inspection date on your report. Then move quickly.

Most suppliers and most payment platforms expect defects to be raised within a short inspection window (often days, not weeks).

Sit on the boxes for a month, and the supplier gets a fair counter-argument:

You accepted the goods, and whatever happened since is on you.

Inspect on arrival, document immediately, and open the conversation while the shipment is fresh.

Put it in a one-page defect report:

Loose photos and a rambling message get skimmed and forgotten.

One structured report gets forwarded to the factory owner and acted on.

Assemble everything into a single document; this is the format that gets taken seriously:

• Header: your brand, PO number, batch/lot number, order quantity, date received, date inspected.
  

  ---

• Inspection basis: “Inspected [X] units of [Y] total” and the standard you checked against (tech pack v[X], approved PP sample dated [X], AQL 2.5 major / 4.0 minor).
  

  ---

• Findings table: each defect type as a row description, category (critical/major/minor), count, and the spec or sample point it violates.
  

  ---

• The headline number: total defective units and the defect rate, stated plainly.
  

  ---

• Evidence: numbered photos, each captioned to a row in the findings table.
  

  ---

• Requested remedy: the specific outcome you want and a one-line reason it’s fair (covered in the next two chapters).

That’s it: factual, numbered, and unemotional.

A factory owner reading it can verify each line against their own records, which is exactly what you want.

A report they can check is a report they can say yes to.

Only once you have photos, counts, a defect rate, and a spec comparison in one factual report do you contact the supplier.

Now you’re not complaining. You’re presenting a case.

02:

Your four remedies: rework, discount, replacement, and refund

Founders panic because they think there’s one outcome: get my money back, and it feels impossible.

In reality, you have four distinct remedies, and picking the right one for this defect is what separates a clean recovery from a stalemate.

Choosing “refund” reflexively when “rework” was the obvious fix is how founders end up with nothing.

The four remedies at a glance:

03:

Negotiating a fair resolution:

The goal here is narrower than “win.”

It’s to reach a resolution that protects your margin without needlessly burning a supplier you may still need.

Founders tend to swing to one of two extremes: rolling over and eating the loss or going nuclear and losing a decent factory over a first, fixable mistake.

Good negotiation lives between those, and it’s mostly about how you open.

Open with the file, not the feelings:

Your first message should read like a claims report, not a complaint.

State the PO, the units inspected, the defect count by category, the defect rate, and the spec each defect violates with photos attached.

Then, crucially, propose the specific remedy you want and why it’s fair to both sides.

Don’t leave it open-ended (“what are you going to do about this?”).

A factory presented with organized evidence and a reasonable ask is far more likely to say yes than one handed a wall of anger and asked to volunteer a solution.

Anchor every number to the spec:

Negotiations drift when they get subjective.

Keep dragging the conversation back to the approved references: the tech pack, the spec sheet, and the signed pre-production sample.

“This is lower quality than I wanted” invites debate. 42 of 150 units measure 2 cm or more outside the graded spec on chest width.

When you argue from the approved standard, you’re not asking the factory to admit they’re bad; you’re pointing at a document you both signed. That’s a much easier yes.

Know your walk-away number before you start:

Decide, privately, what outcome you’ll accept and what you’ll do if they refuse.

That fallback is your leverage: a chargeback against the unpaid balance, a dispute through your payment platform or letter of credit, or simply moving your next (and much larger) order elsewhere.

You negotiate far better when you’re genuinely willing to walk, and factories can feel the difference between a founder who’s bluffing and one who has a real plan B.

Reach for partial acceptance:

The fastest deals are rarely all-or-nothing.

In most defective batches, a chunk of the run is perfectly fine.

Accept the good units, and remedy only the defective ones: rework them, discount them, or have them remade.

This splits the problem into a size everyone can stomach and gets your sellable stock moving immediately.

Insisting the entire order is worthless when 70% is fine just hands the factory an easy counter-argument and slows everything down.

Use the balance and formalize the deal:

If you still owe money, the cleanest resolution is often to net the remedy against the outstanding balance: “We’ll release the final payment, less an agreed X for the defective units.”

If you’re paying through an escrow or trade platform, use its formal dispute channel rather than settling only over chat; the platform can hold funds and mediate.

And whatever you agree on, get it in writing.

A revised PO, or even a clear email, both sides acknowledge, spelling out the remedy, who pays freight, the new timeline, and that this closes the claim.

Verbal deals in international sourcing evaporate the moment there’s a second problem.

What a fair settlement actually looks like:

Abstract advice is easy to nod at and hard to use, so here’s the arithmetic on a realistic case.

Say you ordered 300 hoodies.

On inspection, 210 are clean, 60 have a major defect (print sits low, off-spec against your tech pack), and 30 have a minor flaw (a loose thread, easily reworked).

You still owe a 50% balance.

A founder leading with emotion demands a full refund on all 300, gets refused because the deposit is spent on fabric, and stalls for three weeks with a launch date bleeding out.

A founder leading with the file does this instead:

• Accept the 210 clean units. No argument; they’re fine, and fighting over them only slows the deal.
  

  ---

• Rework the 30 minor units. Cheap and fast; the factory eats a few hours of labor.
  

  ---

• Discount the 60 major units. They’re sellable as B-grade, so negotiate a concession roughly the margin you’ll lose reselling them cheaper, netted against the outstanding balance.
  

  ---

The result: your sellable stock ships on schedule, the factory keeps most of its payment, and you’ve absorbed a defined, survivable hit instead of a stalemate.

Nobody “won,” which is exactly what a fair settlement looks like.

The clean units were never in dispute, the minor ones cost almost nothing to fix, and the real negotiation narrowed to the 60 genuinely wrong units.

That’s the whole trick: shrink the fight to only the part that’s actually broken.

04:

When to walk away vs. repair the relationship:

Once the immediate claim is settled, the harder strategic question lands:

Do you keep working with this factory or find a new one?

Founders get this wrong in both directions, abandoning a good manufacturer over a single honest mistake or clinging to a bad one out of sunk cost and the sheer dread of starting the search over.

The switching cost is real, but it’s not a reason to stay somewhere that will keep costing you. Run the decision through four questions.

01. Is this a one-off or a pattern?

A first serious defect after several clean runs is a repairable event.

Every factory has a bad batch.

A pattern: the third order in a row that “mostly” matched spec, a slow creep in quality is a structural signal, not an accident.

One mistake is a mistake. A pattern is who they are.

02. Did they own it, or deflect?

This is the single best predictor of whether the relationship is worth saving.

A factory that inspects your evidence, acknowledges the error, and proposes a fair fix is showing you exactly how they’ll behave next time something goes wrong, and something always eventually does.

A factory that stonewalls, blames you, blames the freight, or goes silent is telling you the same thing.

Accountability under pressure is the whole game.

03. How severe was it, and how are they resolving it?

Weigh the damage against the response.

A major defect met with an immediate, generous remedy can leave you with more confidence than you started with.

A minor defect met with weeks of excuses tells you what a real problem would look like. Judge the resolution, not just the defect.

04. How was communication before the problem?

Very often the defect is a symptom, and the real disease is communication.

Did they go dark during production?

Miss updates?

Dodge questions about the sample?

If the relationship was already frustrating before the boxes arrived, the defect is just the moment it became visible.

Trust your read on the whole run, not only its ending.

Weigh all of this against the genuine cost of switching.

Finding, vetting, sampling, and building trust with a new manufacturer costs months and money; this is exactly why how you vet a manufacturer in the first place matters so much and why a factory that owns its mistakes is an asset worth protecting.

Don’t torch a good relationship over one honest error.

But equally, don’t let the fear of starting over trap you with a supplier who has shown you, more than once, exactly who they are.

05:

Preventing repeats with better QC:

The best defective-order dispute is the one that never happens.

Almost every problem in the chapters above traces back to a gap that was preventable at a specific, cheap moment earlier in the process.

Quality control isn’t a single inspection at the end; it’s a series of checkpoints, and each one you add dramatically shrinks the odds of opening a bad box.

Here’s the sequence, from cheapest-to-fix to most expensive-if-missed.

Start with a tech pack that leaves nothing to interpretation:

Ambiguity is where defects are born.

If your tech pack and spec sheet leave any figure open, such as fabric weight, exact color reference, stitch type, print placement, or graded measurements with tolerances, the factory will fill that gap with whatever is fastest and cheapest for them, and they won’t be wrong to.

A tight, unambiguous tech pack is the single highest-leverage document in your entire production process.

Most “the factory got it wrong” stories are actually “the tech pack didn’t say.”

Never approve bulk without a signed pre-production sample:

The pre-production (PP) sample is your reference standard for the entire run.

You sign it, the factory keeps a countersigned copy, and every unit gets judged against it.

Approving bulk production off a photo, a promise, or last season’s sample is how founders end up with a run they never actually agreed to.

If the PP sample is wrong, you catch it for the cost of one garment.

If you skip it, you catch it for the cost of the whole order.

Inspect during production, not just after:

An in-line inspection partway through the run catches a systemic problem, a misset machine, a wrong trim, or a color drift while there’s still time and material to correct it.

Catching the same issue only at the end means the entire quantity is already made wrong.

Even a simple mid-run photo request from the factory floor can save an entire batch.

Set an AQL and run a pre-shipment inspection:

Before the goods leave the factory, a pre-shipment inspection (PSI) checks a statistically valid sample against your defined AQL.

State the AQL levels directly in your purchase order; apparel is commonly inspected at 2.5 major / 4.0 minor, so “pass” and “fail” are objective, agreed numbers rather than opinions after the fact.

For meaningful orders, a third-party inspection service (firms like QIMA, SGS, Bureau Veritas, or Intertek) can do this on your behalf, so you’re not relying solely on the factory grading its own homework.

The inspection fee is trivial next to the cost of a failed run you only discover after paying in full.

Tie the final payment to inspection:

Everything above only has teeth if your money is still on the table when the goods are checked.

Structure payment so the final balance is released only after the order passes inspection.

This single term converts every checkpoint from a polite request into a real condition, and it’s the mechanism that keeps you out of the weak, already-paid-in-full negotiations from Chapter 3.

Write a PO that actually protects you:

Put it all in one document the factory acknowledges before production starts: full specs referencing the approved sample, the AQL standard, the inspection requirement, the payment schedule tied to that inspection, the timeline, and a plain remedy clause for defects.

It also pays to be precise about shipping responsibility; the ICC’s Incoterms 2020 rules define exactly where the factory’s responsibility ends, and yours begins, which matters enormously when you’re arguing about who owns damage and freight.

A clear PO isn’t bureaucracy.

It’s the document you’ll be grateful for the next time a box opens wrong.

Frequently asked questions