Trade desk: samples, specs and project reviewRequest Samples
2026-11-10 by Dalia Nasser

How Strict Should Your Spec Review Be? A Scenario-by-Scenario Guide

There is no single right level of scrutiny

I get some version of this question maybe three or four times a month: how deep should our spec review go before we approve a batch? The honest answer is that it depends — but not in the hand-wavy, punt-it-back-to-you way that phrase usually means.

Over about four years of reviewing the things that actually reach customers, and somewhere around 200 SKUs a year in our Q1 2024 audit alone, I've landed on this: the right intensity of verification comes down to two things — what it costs when something fails, and whether the customer can actually tell. Map those two against each other and most orders fall into one of four situations, each with a very different level of scrutiny attached.

Here's the breakdown.

Scene 1: High-visibility performance outerwear

This is the one people over-worry about, but they over-worry in the wrong place.

Take Gore-Tex shells — collab drops, limited colorways, ski and snowboard pieces. The common mistake is that the vendor sends a hydrostatic head report on the membrane itself, everybody nods, and it ships. But the membrane is almost never where it fails. It fails at the seams, the tape, the zipper chain, and whatever's left of the DWR after a dozen washes.

It took me about three years and something like 150 sample batches to understand that the waterproofing on a shell is rarely a membrane problem — it's an assembly problem.

In our Q1 2024 audit we caught a batch of 800 units that looked completely fine on paper. Membrane data was clean. We pulled seam samples, tested, and got bubbling at the seam centers under pressure. The vendor's response was 'the membrane passed.' Right — but nobody's selling the membrane. They're selling the jacket. We rejected the batch and they redid it on their dime.

If you're evaluating unisex Gore-Tex ski jackets or anything laminated, the testing you actually want is hydrostatic head at the seams (AATCC 127 or ISO 811 — ask which, because the two aren't directly interchangeable), tape adhesion after washing, and water repellency after a minimum of three home launderings. Put all three in writing.

Scene 2: Everyday apparel

Then there's the category where we were quietly spending way too much.

Polyester dress shirts are a good example. We used to send a lab test for fiber content on every quarterly run. It's not expensive per test, but we were burning hours chasing certificates and holding up approvals for something I later realized no customer had ever complained about.

When I compared two years of quality complaints side by side — same product, same suppliers, different specs — the picture was pretty blunt: almost none of them were about fiber content. They were about color variation between dye lots (ugh, again), shrinkage after the first wash, and crooked stitching. So we moved the testing budget to lot-by-lot shade card comparison and a single pre/post-wash measurement. Complaint rate went down. Testing spend went down. Both at once.

And another thing — if you're ordering 300 or 500 units and someone tells you 'small runs can't get shade cards,' that's a brush-off. Shade cards cost next to nothing and you can get them at much smaller volumes than that. So glad I pushed back on this the first time it came up; almost signed off on a first delivery blind, which would have meant shipping a color that didn't match the one we'd approved.

Scene 3: Upholstery and interior fabric

People ask me what is a good rub count for upholstery fabric almost every week, and my first move is always a follow-up question: which test are you quoting?

Wyzenbeek and Martindale are not measuring the same thing, and the numbers are not comparable. As a rough rule of thumb — and I mean rough, not gospel — for residential use you're looking at 15,000 cycles or so on Wyzenbeek (some labs report that as about 30,000 double rubs) or 20,000 to 25,000 on Martindale. Lower is workable in light-use areas. Commercial use pushes higher, and the specific setting matters more than the blanket number.

Here's the counterintuitive part: more abrasion cycles is not automatically better. A 50,000-cycle fabric built from fine, loosely woven yarns will pill and snag and look tired sooner than a well-constructed 20,000-cycle fabric. Abrasion resistance is one axis. We also look at pilling (ASTM D4970), yarn slippage, and colorfastness — light (AATCC 16) and crocking (AATCC 8).

So when somebody hands you a number, ask what test produced it. If they can't tell you, the number is marketing, not a spec.

Scene 4: Everyday consumables

Brown paper towel is the category everyone assumes needs no spec review. It needs a different kind, not none.

Sourcing usually pins it to basis weight and unit price. Fine, but the thing that actually drives complaints — a box of towel coming apart, a janitorial crew refusing to use it — is wet tensile strength. A towel that holds up dry and disintegrates on a small spill is a bad towel. So that specification needs to sit right next to basis weight: wet tensile strength (referenced to the pull-test method your lab runs it against) and actual absorbency, tested, not the absorbency number printed on the marketing sheet.

These tests are cheap. The math on a 40,000-pack run says your savings from catching one bad batch will be way bigger than the two-point discount you fought for in the price negotiation.

How to tell which scene you're actually in

Ask yourself three questions, in order:

  1. If this thing fails in a customer's hands, will they blame your brand? Yes → Scene 1. Test hard, and test at the assembly level, not the component level.
  2. Is the failure visible but not catastrophic, and are you buying at volume? Yes → Scene 2 or 3. Focus on lot consistency and the specific attributes customers actually notice, not the ones the cert sheet covers.
  3. Is the unit price low but the item high-turn and high-visibility in use? Yes → Scene 4. Lock the one or two measurements that drive complaint rates and leave the rest alone.

If the answer to all three is no, you're probably over-verifying.

One last thing about small orders, because it comes up and it bugs me. When I was starting out, the vendors who treated a 200-unit trial run with the same seriousness as a 20,000-unit run were the ones I was still buying from three years later. Spec review for a small buyer isn't a courtesy — the color card, the seam pull test, the one washed sample, none of it costs much. What's expensive is losing the account when the first batch goes wrong. Small doesn't mean unimportant. It usually means potential.

Dalia Nasser

Dalia Nasser is a textile fibre and yarn analyst specializing in cotton, wool, polyester, nylon, viscose, lyocell, elastane, recycled filaments, staple fibres, technical yarns, and sewing threads. She applies ISO 2061 twist measurement and the ISO 1833 series for blend analysis while comparing tex or denier, tenacity, elongation, evenness, twist direction, moisture regain, filament count, dye uptake, contamination, and lot consistency. Her sourcing notes help spinners, weavers, knitters, and product teams translate fibre declarations and yarn specifications into stable processing and fabric performance.