What Is Cone Denim — and the 4 Fabric Sourcing Mistakes That Cost Me $12,250

Most B2B fabric sourcing failures aren't price failures. They're spec failures — one ambiguous line in a document nobody re-read before the PO went out, which the mill then interprets in whichever way is easiest for them to produce.

I know because it's happened to me four times, and I kept the numbers. Over seven years sourcing denim, rayon, poplin, and polyester programs, I've personally cost my employer about $12,250. Only one of those four mistakes was really about money. The other three were about words.

If you remember one thing: don't write a fabric spec in language a mill can argue with. “Cone denim,” “60s poplin,” “bamboo rayon,” and “recycled polyester” all have two or more valid readings — and the mill's reading is the one that shows up on the truck.

Here's the compressed version of all four, before the stories:

• “Cone denim” is not a fabric type — it's a mill heritage. Put it on a PO without a construction spec and you'll get the look, not the behavior. Cost: $2,400.
• A cheaper rayon distributor with no lot traceability. Saved $340, spent $3,150.
• “60s poplin” without a yarn system. Ne and Nm are 1.693× apart. Cost: $2,900.
• Polyester compliance paperwork treated as a shipping detail. Cost: $3,800.

Why you should take this from me and not from a spec sheet

I coordinate fabric sourcing and OEM/ODM orders for a mid-size textile business — denim bottoms, woven shirting, and a small knit line. I've been doing it since 2019. I'm not a textile engineer. I'm the person who places the order, chases the lab dips, and finds out eight weeks later whether the words I used meant what I thought they meant.

Four significant errors between 2021 and Q1 2024, all documented, all with a dollar figure attached. After the fourth, I stopped trusting memory and built a pre-PO checklist that lives in our shared drive. We formalized it in Q4 2024. Since then it's flagged 58 spec problems before they turned into purchase orders (as of April 2026).

I'm writing it out because most sourcing guides get written by people who've never had to explain a write-off to a CFO. I've done it twice, and the second time I'd run out of excuses.

Mistake #1: I put “cone denim” on a purchase order

This is the one that taught me what “cone denim” actually means, and it's the question I get asked most, so let's answer it properly first.

Cone Denim is a company, not a construction. The denim house was founded in Greensboro, North Carolina in 1891, and for most of the 20th century its White Oak plant was the mill behind classic American selvedge denim — woven on narrow shuttle looms at roughly 29–32 inches wide, with a genuine finished edge. White Oak closed in December 2017. Cone Denim still makes denim, but from mills in Mexico and China, not from those looms.

So when someone says “I want cone denim fabric,” they usually mean one of three things, and only one of them is a spec:

1. A current collection from the actual mill, which you have to specify by collection name and construction;
2. The vintage American selvedge look — narrow width, shuttle-loom irregularity, visible slub, a finished edge;
3. Whatever a sales rep decided “cone denim” means when they quoted you.

My version of this mistake (2021): I quoted a customer on option 2 and ordered option 3. The PO said “cone denim fabric, 12.5 oz, raw, selvedge.” I figured the mill would fill in the rest, because that's what you usually do with poplin or twill. What arrived hit 12.5 oz and matched the color standard — and had a faux selvedge, a folded and stitched edge on wide projectile-loom fabric. Weight: correct. Fade behavior: nothing like the reference pair.

Cost: $2,400 in dead stock, plus a customer who now double-checks everything I send her. Which, honestly, is fair.

What goes on the PO now, every time: loom type, width, selvedge construction (true shuttle vs. faux), warp and weft yarn counts, and whether the indigo is rope-dyed or slasher-dyed. The words “cone denim” don't appear unless a construction spec sits right next to them.

Mistake #2: The rayon distributor that saved me $340 and cost me $3,150

Rayon is the fiber I'd nominate as the most under-specified in the industry. Viscose is a regenerated cellulose fiber, and its behavior — shrinkage, dimensional stability, hand after washing — depends heavily on the process and the lot. Two rolls that pass the same catalog sheet can behave like two different fabrics in a wash test.

So of course I bought the cheap stock lot. A distributor was quoting about $0.31/m under our usual source, and on 1,100 meters that's $341 saved. Looked like an easy win at the time.

The lot shrank 6.4% on the first wash against a 3% allowance, and the dye lot was visibly off from the standard we'd approved three months earlier. Two hundred and ten pieces out of a 400-piece order had to be redone or credited. Net loss: $3,150 against $340 saved. That's the most expensive $340 I've ever found.

Then there was the label. The distributor listed the fabric as “bamboo rayon.” Per FTC guidance on bamboo textiles (ftc.gov), textile products made from bamboo through a viscose process have to be labeled rayon — the fiber is chemically the same as other viscose, and calling it bamboo implies a mechanical process that isn't what happened. The FTC has brought enforcement actions against major retailers over exactly this. If your hangtag says “bamboo” and your fiber-content line says “rayon,” someone in compliance will notice.

Three things I now require from any rayon fabric distributor before the first order: a shrinkage test run on the actual lot (not the catalog), a dye-lot number with a retained sample of the lot I approved, and the fiber content written exactly as it needs to appear on the label.

Mistake #3: “60s poplin” — my words, not theirs

Poplin fabric OEM is where I expected the least ambiguity, and it's where I got burned by the most boring possible thing: a missing unit.

Poplin is a plain weave with a fine warp and a heavier weft, which gives it that faint crosswise rib and the crisp, slightly dry hand shirting customers either love or reject on sight. A lot of suppliers use the word interchangeably with broadcloth, and they aren't the same fabric — its own problem, but not the one that got me.

The one that got me: I wrote “60s × 60s poplin, 120 gsm” on a spec sheet. The mill read “60” as Nm. I meant Ne. Those are 1.693× apart — Nm 60 is roughly equivalent to Ne 35. So instead of the 110–115 gsm shirting I'd sampled, I got something in the 140 gsm range.

On a swatch, that reads as “a touch heavier, not bad.” On 1,800 meters and 900 finished shirts, it reads as a rejected hand, a chest-print registration problem, and a CMT cost increase nobody approved. A missing unit is a deal-breaker, and it's completely invisible until the fabric is on your floor.

Cost: $2,900 plus a three-week delay that cost us a photo shoot slot. All of it from one letter that didn't get written.

The spec sheet now carries yarn count with the system (Ne, spelled out), measured yarn count, ends per inch and picks per inch, and shrinkage allowance with the test method. Three of those four fields have caught real errors since.

Mistake #4: Treating polyester compliance paperwork as a shipping detail

Polyester is where compliance stops being a formality and starts being a line item. It's also the fiber with the most vague marketing language orbiting it — recycled, eco, sustainable, moisture-wicking.

In Q1 2024 we shipped a polyester program to a customer whose retail partner wanted a recycled polyester claim on the hangtag. I'd confirmed the yarn was rPET. What was still up in the air was the documentation chain — whether the recycled content was backed by traceability records the retailer would accept — and nobody had pulled the current restricted-substances evidence for the dyed trim.

The shipment held at the DC. We air-freighted a partial to keep the launch date. Cost: $3,800.

Two things I learned the expensive way:

• Recycled-content claims need substantiation, not assurance. Per the FTC Green Guides (16 CFR Part 260, ftc.gov/green-guides), recycled-content claims have to be backed by evidence, and “recycled,” “pre-consumer,” and “post-consumer” aren't interchangeable terms. A distributor saying “yeah, it's recycled” on a call isn't substantiation. If you can't document the content, that claim is a red flag for your customer's legal team, not a selling point.
• Chemical compliance is a document set, not a checkbox. Depending on market and end use, that set can include OEKO-TEX Standard 100 test reports, EU REACH restricted-substances declarations (azo dyes under Annex XVII are the usual tripwire on dyed polyester and trims), CPSIA requirements for US children's products, and California Prop 65 sitting on top for US retail.

I'm not going to tell you a fabric “passes compliance.” Nobody can promise that for your end use and your market — the requirements shift by destination, by age group, and by retailer, and if a supplier tells you everything's covered, that's your first warning sign, not your last.

Four questions go on the spec sheet now, before the PO: which test standard and what's the report date, which destination market, adult or children's end use, and who owns the label. Cheapest part of sourcing, and it's where the big surprises hide.

Why the fix was process, not expertise

Here's what connects all four.

After mistake #4 I went back through two years of files. When I compared our 2022–2023 orders — the ones running on email threads, Word docs, and a spec sheet that changed depending on who typed it — against orders that went through a single standardized template in 2025, the mills weren't different. Plus, the fabric categories weren't different. The difference was that in 2025 every field was mandatory and every number had a unit attached.

That's basically it. That's the whole fix, and it's not glamorous. Bottom line: the variable wasn't the mill. It was the paperwork.

We run a hard gate before a PO can be issued. No PO until these are filled: yarn count with the system, fiber content exactly as it appears on the label, width and loom type for denim, shrinkage allowance with test method, dye method, and the compliance document set for the destination market.

Since we formalized it in Q4 2024, the checklist has flagged 58 spec problems before they became purchase orders (as of April 2026). Three of them were five-figure problems in the making. The rest were small. Small ones are how the five-figure ones start.

Honest caveat: this isn't smarter than a good merchandiser. It just doesn't get tired at 6 p.m. on a Friday, and it doesn't assume the mill knows what I meant. Traditional hand-to-hand sourcing still wins on anything genuinely new — the process is best at the boring, repeatable work, which is most of what a sourcing calendar actually is.

Where this doesn't work

A few boundary conditions, because a checklist applied everywhere ends up applied nowhere:

• Development-stage fabrics. A pre-PO checklist is a pre-PO tool. It can't validate a construction that doesn't exist yet.
• Very small runs. Under a few hundred meters, a full test package often costs more than the order. For sampling and one-offs, a retained sample plus a garment-level wash test is usually the better trade.
• Novelty and specialty finishes. Coated, bonded, and heavily finished fabrics behave in ways a spec sheet doesn't capture. Those still need a physical approval round with hands on the goods.
• Anyone else's test data. Third-party reports are tied to a specific lot and date. If the lot changed, the report doesn't cover it. Ask for the lot number, not the report.
• Your own team. A checklist nobody fills in is decoration. Ours only works because the PO literally can't go out without it.

One last boundary: prices, lead times, and mill availability move constantly. Everything above is about process, which is the part that doesn't need re-quoting every quarter.

I still make mistakes. They're just cheaper ones now. If you're working through a denim, rayon, poplin, or polyester program and want the checklist our sourcing team at cone-denim runs, ask for it — it gets revised every time it fails us, which is roughly twice a year.