Last Q1, we rejected a roll of awning fabric that looked fine on the invoice. The color was off by a shade you'd only catch next to the approved swatch. The hand felt stiffer than the sample. And the supplier gave me the answer I've heard maybe fifty times: “That's within industry standard.”
Here's what four years of reviewing first articles has taught me: most of the time, the supplier isn't lying. Your spec sheet is. Or more precisely, it's silent on the one thing that matters—what “standard” means in numbers.
If you've ever opened a roll and known something was wrong before you could name it, you know the feeling. You're not imagining it. But the fix usually isn't a new supplier. It's a better definition.
“Standard” is a word, not a number
The first thing I ask when a supplier says “within standard” is: which standard, and what's the tolerance? Nine times out of ten, the answer is a shrug dressed up as confidence.
Color is the clearest example. We specify brand-critical colors against Pantone references with a Delta E tolerance. Delta E under 2 is roughly imperceptible to most people; 2–4 is noticeable to a trained eye; above 4 is visible to almost anyone. If your PO says “match Pantone 286 C” but doesn't say Delta E ≤ 2, you've given the supplier room to deliver something that technically passes and still looks wrong on the shelf.
I learned this the hard way in 2022. I said “match the previous batch.” They heard “match the Pantone number on the PO.” Result: a batch that hit the Pantone but missed the hand of the original—same color family, different drape, and a customer who noticed on day one. Nobody was dishonest. We were using the same words and meaning different things. We only discovered it when the rolls arrived and the upholstery team said the fabric “felt like a different product.”
Now every PO carries the reference, the tolerance, and the test method. It's three extra lines. It has saved us more arguments than any supplier negotiation.
The fiber-name problem (this is where lyocell and viscose get messy)
Color is easy compared with fiber content. Ask ten buyers what is viscose fabric and you'll get ten variations of “a soft plant-based fiber.” That's not wrong—viscose is a regenerated cellulose fiber, usually from wood pulp, made through a chemical process that's been around for over a century. Soft, breathable, drapes well. Also prone to shrinking and weaker when wet.
Then you get lyocell fabric, which is also regenerated cellulose—but made with a different solvent system (NMMO) in a closed loop. The lyocell fabric characteristics people actually care about are different: higher wet strength, better dimensional stability, a smoother surface, and a much lower environmental load from the solvent recovery. Same family tree. Different animal.
Here's where it bites you. A label says “lyocell blend.” No percentage. No test data. The supplier technically told the truth. But if you assumed 60% lyocell and got 15%, you're not buying the same performance—you're buying a story. On a curtain or contract upholstery program, that gap shows up as shrinkage after the first cleaning cycle, or a hand that doesn't match the sample book.
I don't have hard data on how often fiber blends get labeled loosely across the industry—my sense from our own intake is that it's common, maybe a third of the blended substrates we see. What I can tell you is that “lyocell” on a spec sheet without a percentage and a test method is a marketing word, not a specification.
And yes, I've been fooled by this too. In 2023 I approved a blend because the swatch felt right and the supplier's email said “high lyocell content.” We found out at the first wash test. Not catastrophic—but expensive enough that I now require the fiber percentage on the lab dip sheet, not just the invoice.
“Water-resistant” and “eco-friendly” are not specifications
Performance claims have the same problem, just with more legal exposure. “Water-resistant” on a spec sheet could mean anything from a light spray rating to a coating that survives twenty washes. AATCC and ISO test methods don't always produce the same rating on the same piece of cloth. If your spec doesn't name the method and the pass threshold, you're not specifying performance—you're specifying hope.
Environmental language is worse. Per the FTC Green Guides (16 CFR Part 260), a “recyclable” claim should hold in areas where at least 60% of consumers have access to recycling. “Eco-friendly” isn't defined at all. When a supplier sends me a tech sheet with a leaf icon and no data, I treat it as decoration until they send the underlying test report.
“If you can't test it, you can't specify it. If you can't specify it, you're negotiating after the fact—and you've already lost the leverage.”
That's the line I use with our own product team. It sounds harsh. It's also the reason our sample approval form now has a column for “claim + test method + threshold” instead of just “claim.”
What this actually costs when you don't catch it
The invoice price is never the real price. The real price includes the redo, the expedited freight, the customer conversation, and the quiet hit to your brand when someone notices the second batch doesn't match the first.
A few years back we had a batch where the spec was visibly off—about 8% shrinkage against our 3% limit. Normal tolerance for that substrate was 3–4%. The vendor claimed it was “within industry standard.” We rejected it, and they redid it at their cost. But the redo still cost us around $22,000 in expedited freight and lost production time. (Should mention: most of that number was freight, not the fabric itself.)
Then there's the storage problem. We once had 8,000 units sit in a warehouse for six weeks while we argued about a coating spec. By the time we resolved it, the edges had taken on moisture and the whole lot was compromised. Nobody's fault, technically. Everybody's problem, commercially.
The part that stings most is the customer relationship. Your buyer doesn't blame the mill. They blame you. And they're right to—you're the one who signed the PO.
How I fixed our spec sheet (short version)
I'm not going to pretend this is complicated. It's just tedious, and most teams skip it because it feels like paperwork.
Four things go on every fabric PO now:
- Fiber content by percentage, not just fiber names. “Lyocell/polyester blend” is not a spec. “55% lyocell / 45% polyester, ±3%” is.
- The test method and the pass threshold, written next to every performance claim. If it's water repellency, name the method and the minimum rating. If it's color, name the Pantone reference and the Delta E limit.
- A first-article requirement with a signed test report from a third-party lab, not a supplier-generated PDF.
- A line that says what happens if the first article fails. Who pays for the redo, who pays for freight, and how fast. Vague consequences are the same as no consequences.
When we started specifying awning fabric for a large outdoor program, we benchmarked our requirements against published technical data—Dickson awning fabric documentation among the references we used—not because a data sheet guarantees performance, but because a good one tells you exactly which claims are testable and which are just words. That's the whole point. A Dickson fabric spec sheet without a tolerance is a brochure. With one, it's a contract.
The uncomfortable truth is that transparency cuts both ways. If you want your suppliers to be clear about what they're delivering, you have to be clear about what you're asking for. The vendors who list every tolerance and test method up front usually look more expensive on the quote. In my experience, they're cheaper by the time the roll is on the floor.
I wish I'd tracked rejection rates more carefully from the start. What I can say anecdotally is that since we rewrote the spec template in 2022, our first-article rejections dropped by roughly half. Maybe 45%—no, closer to 50%, I'd have to pull the log to be exact.
The number I care about more: the arguments after delivery. Those went to almost zero. Turns out most of them were never about quality. They were about definitions we never wrote down.