TL;DR: A supplier’s Certificate of Analysis tells you more about their production discipline than their sales pitch ever will — learn what fields to require and what gaps to reject.
TL;DR: In our incoming inspection protocol, we flag any COA missing burst strength, caliper, or GSM values as a Category B hold — roughly 1 in 8 lots from new suppliers fails this screen on first submission.
What a Complete COA Should Contain — and What Missing Fields Actually Mean #
When a new material supplier sends us a Certificate of Analysis, the first thing we check is not the numbers — it’s which fields are present. A COA with values for every parameter tells us the mill ran a full test sequence. A COA with three fields completed and seven blank tells us they tested selectively, which means they already know which results don’t look good.
For paperboard and greyboard (our highest-volume incoming materials), we require a minimum of 12 fields on every COA submission:
| Parameter | Acceptable Range | Test Method |
|---|---|---|
| Grammage (GSM) | ±5% of nominal | ISO 536 |
| Caliper / Thickness | ±7% of nominal | ISO 534 |
| Burst Strength | ≥ 300 kPa for 300gsm SBS | ISO 2759 |
| Moisture Content | 6–9% | ISO 287 |
| Whiteness (if coated) | ≥ 82 CIE | ISO 11475 |
| Stiffness (MD/CD) | Per grade spec | ISO 2493 |
Any lot arriving without burst strength and caliper values goes into our QC-Hold-M2 queue pending re-test at our incoming dock. We do not accept “values available on request” language — if the mill cannot include the data in the shipment document, that is a procedural gap worth probing before you commit to a 50,000-unit run.
The burst strength threshold matters especially for shipping cartons destined for the US or EU. Carriers operating under ISTA 2A and ASTM D4169 drop-and-vibration sequences will stress carton walls in ways that reveal under-weight board early. A grammage variance of just 8% below nominal (outside our ±5% acceptance limit) can reduce stacking strength by 12–15% — enough to cause pallet collapse at a 3PL facility.
Where Qualification Breaks Down — Three Failure Modes We See Repeatedly #
The most common failure mode in supplier qualification is treating a COA review as a one-time event at onboarding, then never repeating it. We run lot-by-lot COA review for every incoming shipment from suppliers in their first 12 months with us, then move qualifying suppliers to quarterly audit mode under our AVL Gate Review procedure. Suppliers who shortcut this sequence often deliver consistent quality for the first two or three lots — the ones they know will be inspected — then drift on lot four onward.
The second failure mode is ignoring ink and coating chemistry documentation. Many brand buyers focus entirely on structural substrates and treat ink supplier credentials as a converter’s internal matter. For food-adjacent packaging — outer cartons for tea, coffee, confectionery — this creates real compliance exposure. Primary food contact packaging in the EU falls under Regulation (EC) No 1935/2004 and, for plastics, EU 10/2011. Paperboard outer packaging is regulated at the national level across EU member states, with Germany’s BfR recommendations and France’s DGCCRF guidelines being the most demanding. We ask every ink supplier to provide a compositional declaration and a migration statement compliant with GB/T 10004 for domestic market jobs, and a full Cramer classification summary for EU-bound jobs. A supplier who cannot provide this within five business days of request is, in our experience, unlikely to maintain the documentation trail you’ll need if a retailer compliance audit ever arrives.
The third failure mode is conflating price with qualification status. We have logged this pattern repeatedly in our Category B incident tracker: a supplier passes initial qualification, then offers a revised price 8–12% below the original quote six months later. When we re-test the material, caliper is typically down 0.05–0.10mm from the original certified value. The supplier has reformulated or changed their furnish mix to hit the new price point. The cost saving is real. The risk is that your printed pack now has a different feel, different rigidity, and potentially different print holdout — which means your brand’s colour density and gloss profile will shift even if nothing changes on the press.
Does Supplier Certification (ISO, FSC, BRC) Replace a COA and Incoming Inspection? #
No. Certification confirms a supplier has a management system in place — it does not certify the specific lot of material sitting on your loading dock.
ISO 9001:2015 certification tells you a supplier has documented procedures and internal audits. FSC Chain of Custody certification (required if you are making FSC-labelled packaging) confirms fibre traceability but says nothing about caliper, burst, or moisture values for that shipment. BRC/IoP Packaging Certification covers hygiene and food-safety management systems. All three are necessary for a credible supplier relationship. None of them replace lot-level COA review or skip-lot incoming inspection.
For critical structural components — rigid box greyboard, folding carton SBS stock above 350gsm, corrugated medium for export shipping — our practice is to run 100% COA verification plus AQL Level II sampling inspection on every lot, regardless of the supplier’s certification status. For lower-risk items such as tissue wrap or ribbon, we move to AQL Level I once a supplier has passed 6 consecutive lots without deviation.
Specification Notes for Brand Partners #
When you brief us on a new packaging project, the clearest way to speed up supplier qualification and sampling is to share your compliance destination market upfront. “We sell into Whole Foods US and REWE Germany” narrows the regulatory framework immediately — and tells us which COA fields and ink declarations to require from day one, rather than discovering the requirement at pre-shipment inspection.
The brief gap that causes the most sample iterations is leaving surface finishing unspecified. Aqueous coating, soft-touch laminate, and UV spot varnish each require different ink-cure energy levels (typically 80–120 mJ/cm² for UV flexo vs. oven-dwell time for aqueous) and interact differently with your substrate’s moisture content. If you specify the finish late, we may need to re-run material qualification for the coating chemistry.
Our standard sampling timeline for folding carton projects is 18–22 working days from approved dieline and confirmed material specification. Rigid box sampling runs 22–28 working days. These timelines extend if incoming material COA review triggers a hold — which is why early supplier alignment saves clock time, not just money.
Frequently Asked Questions #
If a supplier has ISO 9001 certification, do we still need to review COAs lot by lot?
Yes — ISO 9001:2015 certifies the management system, not the physical properties of each shipment. A certified supplier can still ship an out-of-spec lot, and certification does not create a paper trail for that specific material unless the supplier’s own IQC process is rigorous. Require COAs regardless.
What AQL level should we specify for incoming carton inspection?
It depends on the application and your retailer’s requirements. For cosmetics and premium food packaging destined for EU or US retail, AQL 2.5 under ANSI/ASQ Z1.4 is a common floor. For general shelf goods, AQL 4.0 is more typical. We apply Level II sampling intensity by default, and tighten to Level III if a supplier has had a nonconformance in the prior 90 days. If your retailer has a specific AQL requirement in their vendor manual, that number governs — we’ve seen Walmart US specify AQL 1.5 for certain food-contact secondary packaging, which is tighter than many brands anticipate.
How do we evaluate a new supplier’s COA quality before placing a first order?
Ask for sample COAs from three recent production lots before any commercial discussion. Look for consistency of values across lots, presence of all required fields, and whether the test methods cited match ISO or ASTM standards relevant to your market. A supplier who submits COAs with identical values across multiple lots (not just within tolerance, but literally identical to two decimal places) almost certainly does not run per-lot testing — those are copy-pasted documents, and that is a disqualifying finding under our QC-Hold-M2 criteria.
Planning a packaging project? Contact our team to request a complimentary specification review and sample quote.
Had a mill in Zhejiang send us a COA for 350gsm SBS last year with burst strength listed as “per customer spec” — no actual value, just that phrase. Took two weeks of back-and-forth before they admitted their tester was out of calibration and the lot had been sitting in their warehouse since March. We pulled it from the 40,000-unit run and found a Shandong supplier who could turn around a compliant COA within 48 hours of shipment.
The re-test cost on QC-Hold-M2 lots adds up faster than most people account for — we’re running about $0.04/unit in dock testing labor on flagged lots, and at 50k units that’s $2,000 in sunk cost before you’ve even decided whether to accept or reject the shipment. Qualifying suppliers on COA completeness upfront, not just the numbers, cut our hold rate from roughly 1-in-6 to 1-in-20 within two seasons.
We’ve flagged the same thing — a COA missing burst strength data almost always means the mill ran ISO 2759 and didn’t like what they saw. Caught it twice in Q3 with a greyboard supplier out of eastern Europe before we pulled them from our approved list.
Moisture content is the one that bites us most on West Coast receipts, especially in Q4 — we pulled 3 lots from a BC-based supplier last November where moisture came in at 11.2%, well outside the 6–9% window, and the caliper readings were technically fine, which is exactly why you can’t use caliper alone as a proxy for board condition.
Caliper tolerance was the hidden one for us — had a Guangdong SBS supplier consistently hitting GSM within spec but running 4–6% thin on caliper across three consecutive lots, which only showed up because we cross-reference ISO 534 against our own incoming gauge checks. Their COA listed caliper values that were technically “within ±7%” but always at the low edge, and it wasn’t until we overlaid six months of data that the systematic drift became obvious.