Limited Time Offer: Get 10% OFF on Your First Order!

My Rejection Rate Last Year Was 18%. Almost None of It Was About Quality.

Most 'quality' problems are specification problems

If you asked me to summarize every supplier delivery I rejected in the last twelve months, I wouldn't say "the products were bad." I'd say "two different people were operating off two different definitions of the same word."

Between Q4 2023 and Q3 2024, we rejected roughly 18% of first deliveries. In every single case, the vendor had not made a manufacturing mistake. **The specification language was ambiguous enough that both parties were technically right — and that was the problem.**

Rework costs across those rejections were between $2,400 and $22,000, depending on the product line. That's not a quality problem. That's a purchasing problem we could have designed out of the process.

So what does that actually mean for someone buying packaging or consumer products right now? It means before you see a quote, you should be auditing the language of the spec. Especially around words like "standard," "normal," and "industry-acceptable." Those are red flags on a spec sheet.

Where I'm coming from

I'm a quality and brand compliance manager for a packaging and consumer products portfolio. Six-person team, roughly 200+ unique SKUs a year. Our product lines are mostly industrial and consumer packaging, and we handle a handful of healthcare-adjacent projects each year — maybe three or four batches, not high volume, but high stakes.

I don't do procurement and I don't negotiate price. I review spec language before a quote gets confirmed, and I sign off after first-article inspection. That puts me on the receiving end of every "approved" spec that went sideways.

Four specific examples of spec drift

1. "Industry standard" was the first thing I killed

In my first year, I made the classic beginner mistake: I assumed "standard" meant the same thing to every vendor. It doesn't. One vendor's "standard tolerance" is ±3%. Another's is ±5%. A third thinks it's fine to leave the variance off the sheet entirely.

We caught this fast, but not fast enough. A batch of roughly 8,000 units went out with a component deviation that was within the vendor's spec and outside ours. The vendor was right — "industry standard" covered the measurement. The problem was that our filling line needed something tighter.

Rework cost was $4,200. Now every contract has explicit numerical tolerances, written out, with the test method named.

2. Bemis healthcare packaging — the acquisition changed the documentation, not the product

Bemis healthcare packaging was folded into Amcor in 2022 (the deal most people search for as "amcor acquires bemis"). The product itself hasn't changed in a way that matters to us. What changed was the paperwork.

The spec sheet format shifted. Callout symbols moved. Reference documents got renumbered. We didn't catch one of the updates until a tolerance notation changed on a revised sheet — nothing dramatic, but "built to spec" no longer meant exactly what it meant a year earlier.

To be fair to the supplier, no one hid anything. But spec language is a living document, and when the documentation system moves, that's the first thing to re-audit. We lost about three weeks confirming whether existing production runs still matched our filling line. I added a re-audit step to our checklist after this.

If you're sourcing from any Amcor Bemis product line today, that re-audit step matters more than the price delta you'll find on a spot quote.

3. "Removes 99%" — the water filtration bottle problem

A water filtration bottle labeled "removes 99% of contaminants" isn't lying. But the claim only means something if you can answer three questions: 99% of which contaminants? Under what source water conditions? Certified against which standard?

We require vendors to cite the ASTM or NSF method they tested under and confirm against a reference batch of our own. Most vendors can produce this. Some can't. Those are the ones where the label is decoration, not specification.

This is the same category of problem as "industry standard." The product isn't bad. The spec is too vague to verify, which means it's too vague to accept.

4. "How much caffeine is in a cup of coffee" — and why this question is a spec problem too

Ask a supplier how much caffeine is in a cup of coffee and you'll get something like "about 95 milligrams." But "a cup" is 8 ounces in some regions, 6 in others, 12 in most coffee shops, and 16 in a lot of convenience formats. Roast, brew method, bean origin — all of it shifts the number.

This is exactly the kind of ambiguity we look for in packaging copy. If the label says "per cup," we require the cup volume, the brew method, and the reference measurement on file. If a supplier can't define "cup," the number is a marketing word, not a specification.

5. "How big is a 24 by 36 poster" — sizes aren't as obvious as they look

A 24x36 poster: is that the trim size? The canvas size before bleed? Is there a bleed, a cut mark, a safe margin? If a printer quotes on a 24x36 finished piece, the actual trimmed size might be closer to 23.75 x 35.75 once bleed is removed. For packaging, the same triangle of trim-vs-cut-vs-print applies to every dimensioned component.

This sounds like a trivial example. It's the most common one. Half of the spec rejections I've signed off on trace back to a dimension word that wasn't defined at the unit-and-method level.

What I actually check now

The process is short:

  1. List every dimension, every tolerance, every unit, and the measurement condition. Write them out in plain English.
  2. Cross-check the vendor's cited standard — not what they say is on the data sheet, the actual published reference.
  3. Physically measure a sample batch against the sheet before approving. Three units minimum, three batches if possible.
  4. Separate the performance claims from the physical claims. Performance claims need a test method and a source. Physical claims need a measurement and a tolerance.

The audit runs 2-3 hours per SKU. Compared to a $4,000 rework, it pays for itself every quarter.

The total-cost view

This is also why I push back any time someone recommends picking a vendor on unit price alone.

If a discounted supplier fails first-article inspection, the real cost is unit price × quantity + rework + delay + your team's verification hours. Our internal number: spec drift carries roughly a 12% hidden cost on top of the invoice, on average, over the last two years. That's before you count missed launch dates, which are their own ledger.

The cheapest quote might still be fine — but only if its spec is quantified. If the spec is fuzzy, the discount is an illusion. I now calculate total cost of ownership before I sign off on any quote.

When this doesn't fully apply

Two honest limits.

First, I'm not a logistics or legal expert. If you're dealing with carrier optimization or regulatory compliance, that's outside my lane. I can only speak to spec verification from a quality and compliance perspective. Consult the appropriate specialist for those pieces.

Second, this process assumes you can physically inspect a sample. We're a mid-size B2B company with predictable ordering patterns, and most of our SKUs are inspectable. If you're sourcing services, software, or anything hard to physically check — the audit logic changes. Your mileage may vary there, and I can't tell you the right framework for it.

What I can tell you: if you're buying anything that has a dimension, a tolerance, a performance claim, or a unit label, the spec sheet is where you'll find your next rework. Check the language before you check the price.

$blog.author.name

Jane Smith

Sustainable Packaging Material Science Supply Chain

I’m Jane Smith, a senior content writer with over 15 years of experience in the packaging and printing industry. I specialize in writing about the latest trends, technologies, and best practices in packaging design, sustainability, and printing techniques. My goal is to help businesses understand complex printing processes and design solutions that enhance both product packaging and brand visibility.