What “A-grade” actually means. And who is actually looking at your goods.

There is no grading authority in stocklot. No inspectorate, no certificate, no schedule anyone signs. When a seller writes “100% A-grade, original packing” into an email, that sentence has exactly as much force as the person who typed it. It is a claim, not a specification — and the only thing that ever converts a claim into a fact is somebody standing on the warehouse floor with the carton open and the polybag cut.

So the useful question is never what grade is it. It is who opened it, how many pieces did they open, and what happened to what they found. This piece answers all three for our own goods, and gives you the vocabulary and the arithmetic to put the same three questions to anybody else.

The words, and what they conventionally mean

A-grade. Sale-floor ready. No visible defect at arm's length, correct label, correct measurements, no odour, no storage marking. In a well-run lot this is the bulk of the goods — but “100% A-grade” across a whole leftover lot, offered without a number attached, is a seller telling you they have not counted.

B-grade, or seconds. A real garment with a real fault: a skipped stitch, a shade variation between panels, a small oil mark, a broken zip pull, a mis-set button. Sellable, at a lower price, to a market that expects it. Useful goods — provided somebody told you they were in there.

Mixed, or assorted. An honest word for an unsorted lot, and often a warning that nobody has opened the cartons since they left the factory. The grade distribution inside is unknown to the seller as well as to you. That can be a genuine bargain; it can also be someone else's problem, repackaged.

Export surplus, export quality. Origin, not condition. It tells you the goods were made to an export buyer's specification — which is meaningful, because that specification is usually stricter than a domestic one. It says nothing whatsoever about what has happened to the garment since.

Original packing. Still in the factory's polybag with the factory's carton markings. Genuinely valuable as provenance — and genuinely useless as proof of condition, because a sealed polybag is evidence of a good polybag and nothing else. We will come back to this.

Cut labels, de-branded. The brand identification has been removed. Necessary and normal on a lot released by a brand. But it also removes your ability to confirm what the garment was, so it raises the standard of proof required on everything else.

Who is actually looking at your goods

In most stocklot deals, nobody is — or near enough. A broker agrees the lot on the phone, and when the mill says come and collect, he sends whoever is free that afternoon. A cousin. A friend of a friend. Somebody's driver. A man hired at the gate for the day. Tom, Dick and Harry, and not one of them a checker.

That person has no defect list, no measuring tape, no idea what the buyer's market will reject, and — this is the important one — no authority to refuse anything. He walks the aisles, looks at the tops of a few stacks, counts cartons, signs the challan, and the truck leaves. The inspection that your money depended on took forty minutes.

This is not really a moral failing. It is an incentive structure. A broker earns on paperwork and speed against a fixed commission, so every hour his man spends cutting polybags is an hour of pure cost to him, and every fault that man finds is a negotiation he did not want. He is not paid to find problems. He is paid to move cartons. The goods that turn up at your door are the arithmetic of that sentence.

We do it the other way round for a structural reason, not a sentimental one: we carry the risk he doesn't. When a container lands in Dubai or Santiago or Lagos and the goods are wrong, the broker is already out of the transaction and unreachable. We are not out of it — we are the ones who have to make it right, and we would rather find the problem in Bangladesh where it is cheap to fix than in your warehouse where it isn't.

What a professional apparel inspection actually is

Worth setting out the industry benchmark before we tell you where we depart from it. When a brand buys new production and engages a third-party inspection house — SGS, Intertek, Bureau Veritas — this is roughly the shape of the work, and it is the reference standard the whole trade measures itself against.

The sample is drawn statistically. Under ISO 2859-1 (ANSI/ASQ Z1.4 in the United States), at General Inspection Level II, batch size maps to a sample size on a published table. Fifteen thousand pieces gives a sample of about 315. Pieces are supposed to be pulled at random across cartons, sizes and colourways — never off the top layer, which is the oldest trick on the floor and the first thing a good inspector refuses to do.

Every fault is classified and tolerated separately. Three classes, each with its own Acceptable Quality Level. Critical, AQL 0 — a broken needle left in a garment, a restricted chemical, mislabelled childrenswear: one is enough to fail the lot. Major, AQL 2.5 — broken stitching, wrong measurements, a hole, a missing component; the faults that come back to you as returns. Minor, AQL 4.0 — loose threads, slight shade variation within tolerance, a light crease mark. On a 125-piece sample that arithmetic allows 0 critical, 7 major and 10 minor defects before the shipment is rejected.

And each sampled piece is worked through a fixed list. Style, colour and size against the specification. Every point of measure against its tolerance. Workmanship: seam construction, stitch density, puckering, raw edges, symmetry, run-off. Trims and accessories: zips cycled, buttons and snaps pull-tested. Labelling: size, fibre composition, care instructions, country of origin, barcode scanned. Then the packing itself — polybag, fold, assortment ratio, carton marks, drop test. Fabric, where it is graded as fabric, is scored on the 4-point system under ASTM D5430: one to four penalty points by the length of the flaw, no more than four points charged against any single yard, and a roll passes at somewhere between twenty and forty points per hundred square yards depending on the buyer.

It is a good system, and we borrow its grammar wholesale — the defect classes, the checkpoint list, the discipline of writing everything down. But it is, unavoidably, sampling. Three hundred and fifteen pieces out of fifteen thousand is two per cent. And it was designed for a situation we are never in: one factory, one order, one fabric batch, one production run, where a random two per cent genuinely predicts the other ninety-eight.

Which is why we open all of it

A stocklot exists because something failed. An order cancelled mid-shipment. A shade batch rejected at approval. A delivery that missed its window. A season that closed while the goods were still on the water. Whatever the failure was, it is usually not spread evenly through the lot — it sits concentrated in the specific cartons that caused it. A two per cent random sample can walk straight past those cartons, pass the lot, and be statistically flawless while being commercially worthless.

So we do not sample. Our own team — our people, on our payroll, trained on our defect list, with standing authority to reject — goes to the mill and inspects the lot piece by piece, one hundred per cent, for however long it takes. Three days on a container is ordinary. We have spent a week on one lot and considered it time well spent.

The packaging comes off first, and it goes in the bin. All of it — polybags, tissue, inner cartons, printed inserts. That is deliberate, and it is the single biggest difference between our inspection and a walk-through. Packaging is where the story gets told: it is what makes a stack look untouched, what hides the middle of a pile, and what a seller is relying on when he asks you not to disturb the goods. So we destroy the story before we assess the garment. Every piece comes out, gets opened flat, turned over, and gone through by hand.

What we are hunting for is the third-party list run without a tolerance table: holes, pulls, snags, thread breaks, skipped and run-off stitching, seam puckering, oil and dirt marks, shade variation panel to panel and piece to piece, mildew and storage odour from long warehouse time, measurement drift, cracked prints, misplaced embroidery, wrong or missing labels, dead zips, loose buttons, needle risk. A hand inside every garment. Both sides of every seam. Twenty thousand pieces means twenty thousand pieces looked at, not three hundred and fifteen.

Because mills are cunning

We say that without heat. It is simply the commercial reality of the trade, and anyone honest in Bangladesh's stocklot business will tell you the same thing: a mill sitting on goods it cannot ship has exactly one objective, and your margin is not it.

The methods are consistent enough to list. Good pieces on top, faults in the middle and the bottom of every stack. The bad cartons at the back of the pile, or behind the pile, or in the next room. Two lots quietly married into one so the weak one ships out under the strong one's count. Pieces already rejected once by another buyer, re-polybagged and returned to the offer as fresh. A carton weight that does not reconcile with the piece count. Rejects blended in at a ratio somebody has carefully calculated to survive an AQL sample.

And then the sentence itself, which we have learned to read as a warning rather than a reassurance: the goods are intact, ninety-nine point nine nine per cent fine, no need to open. In our experience that sentence has almost never been true. Whenever we have heard it, something was wrong inside — and the more strenuously a seller argues that we needn't look, the more carefully we look. We have never once regretted the hours.

Then we check it again, at home

Everything we buy is inspected a second time in our own warehouse in Dhaka, by a different team, against the same list. Not because we distrust the first inspection — because a second pass is what keeps the first one honest, and because loading, transit and unloading generate faults of their own.

The interesting part is the result. The second inspection barely deducts anything. Fractions of a per cent, routinely. That is not a boast about our standards; it is the only meaningful audit of them. If the mill-floor inspection is real, the warehouse inspection has nothing left to find. And when a second number does climb, we know precisely which team was on the first job — that is a conversation we have internally, instead of passing it on to you as a surprise.

And the faults never get on the boat

Here is the part that changes the arithmetic completely. Everything the two inspections pull out — the seconds, the marked, the holed, the odd measurement, the pieces we would have had to write you an apologetic email about — is separated and stays in Bangladesh. It is sold locally, at local prices, to recycling mills who pull it back to fibre.

You never receive it. The faulty fraction is removed before the lot is priced and long before it is loaded, so you are not paying ocean freight, duty, clearance and finishing labour on pieces you can never sell. You are not storing them, not sorting them, not discounting them, and not arguing with us about them.

We are not out of pocket either. Rejected garments have a real domestic value as recycling feedstock, which is exactly why we can afford to be strict: pulling a piece out costs us the gap between export price and recycling price, not the whole piece. A broker with no local recycling channel cannot afford to reject anything at all — which is precisely why he doesn't, and why his goods reach you with the decision still un-made.

And none of it becomes waste. It goes back to fibre in the country where it was made, a short truck ride from where it was rejected. We would like to present that as a sustainability policy. It is more honest to call it what it is: the commercial arrangement that makes uncompromising inspection affordable, which happens also to be the right thing to do with a garment nobody should wear.

Cheaper is a claim too

All of this only matters because of one calculation, and it is the most expensive one in the business to get wrong. A lot arrives priced well below the market. The invoice looks like a win. What you have usually bought is not a discount at all — it is an undisclosed dead fraction, sold to you at the price of live goods, by somebody who never opened the cartons and therefore never had to mention it.

Because the arithmetic does not work the way instinct says. You do not pay for the pieces you sell. You pay for every piece in the carton — and then freight on it, duty on it, finishing labour on it, warehouse space for it — whether or not it will ever leave your shelf.

Take ten thousand pieces, two ways. Lot A comes from a lot that was opened piece by piece, with the faults pulled out before loading; it is offered at $2.10 and runs 5% unsaleable by the time you have it on your floor. Lot B is the walk-through lot at $1.55 — twenty-six per cent cheaper on the invoice — and, once opened, runs 28% unsaleable. Landed cost on both, freight and duty and finishing together, is about $0.45 a piece.

Lot A costs you $21,000 for goods and $4,500 to land, so $25,500 for 9,500 saleable pieces: $2.68 each. Lot B costs $15,500 and the same $4,500 to land, so $20,000 for 7,200 saleable pieces: $2.78 each. The cheaper lot is the more expensive lot — and that is before you have paid anyone to sort it, stored 2,800 dead pieces you cannot shift, or handled the returns from the ones that slipped through.

The break-even, in one line

Divide the good fraction of the clean lot by the good fraction of the cheap one. Ninety-five over seventy-two is 1.32. That is the whole answer: a lot running 28% dead has to be thirty-two per cent cheaper than a clean lot merely to draw level. Not ten per cent cheaper. Not fifteen. Thirty-two — and at that price it is not a bargain, it is a wash, with every hour of the sorting work still ahead of you.

Run it the other way and the discipline becomes simple. Before you agree a price, put a number on the dead fraction — the seller's number if somebody counted, your own worst-case estimate if nobody did. Add your landed cost per piece to the goods price. Divide by the good fraction. Compare that figure between lots, never the invoice price. A twenty per cent discount that buys a twenty-five per cent dead rate is not a discount at all.

Five questions, before the price

Was this lot sampled, or opened? A percentage is not an answer to this question. “We checked it” is not an answer either. Ask how many pieces out of how many, and listen for whether the two numbers are the same.

Who did the opening, and are they on your payroll? A trained checker who can be held responsible next month behaves differently from a man hired at the gate that morning. If the seller cannot name the team, there was no team.

Show me the faults you found. Any seller can photograph their best piece. A seller who will photograph their worst one — and tell you how many like it there were — is a seller who has actually looked.

Where are the rejected pieces now? There are only two honest answers. Out of the lot, sold or destroyed locally, with a figure to prove it. Or: still in the lot, and about to be your problem. Settle this in writing before shipment, never after opening.

Why is it this cheap? There is always a reason, and a real one is easy to give: a cancelled order, a late shipment, a closed season, a warehouse being emptied. A seller who cannot name the reason is either not the owner of the goods or not telling you what is inside them.

None of this makes cheap lots bad. Some of the best buying anyone does in this trade is on goods with a double-digit fault rate bought at a price that respects it. The unbuyable lot is the one where nobody will tell you the number — because then the discount is not a discount. It is a bet, and the seller has seen the cards.