Quality · Published Standards

Our QC standards, with the numbers

Every supplier says "strict quality control". Almost none will tell you the pass thresholds. Here are ours: four inspection gates on every order, a signed golden sample before any production, and a laboratory test battery with published numbers — vacuum at -0.08 MPa, tape adhesion with 3M 600 and 810, 48°C and -15°C cycling, 76 cm to 1.5 m drops, 1,500-cycle pump life, colour at ΔE under D65 light. If a supplier will not put figures on their QC, they are describing an intention, not a system.

These are the working standards we apply with partner factories, aligned with common industry and international test methods. Product-specific thresholds are agreed in your specification before production.

The System

Four gates, not one final glance.

Most cross-border packaging disputes trace back to a single cause: the only inspection happened when the goods were already made. Our orders pass four gates, and problems get caught where they are cheap to fix:

GateWhenWhat it catches
IQC — incomingBefore materials and bought-in components enter productionWrong resin grade, off-spec preforms, pumps that fail actuation sampling
IPQC — in processOn the line, at set intervalsWhere compatibility, adhesion and leak problems actually surface — decoration drift, weld-line weakness, seal variation
FQC — finalFinished goods, against the golden sampleColour drift beyond agreed ΔE, fit-as-a-set failures, cosmetic defects by A/B/C surface class
OQC — outgoingPre-shipment, sampled to agreed AQLCarton strength, inner protection, quantity, labelling, documentation

The golden sample rule sits above all four gates: production starts only after you approve a physical sample and both sides sign it. Colour tolerance against that sample is agreed in writing and judged at ΔE under a D65 standard light source — not by eye, not under warehouse lighting. A sealed duplicate is retained for the life of the programme, so your reorders are checked against the same physical reference as your first order.

The Test Battery

The tests, with pass thresholds.

TestMethodTypical pass threshold
Vacuum leakFilled unit in vacuum chamber, lying and inverted-0.08 MPa held 15 min, loss ≤0.5% (volatile formulas); -0.06 MPa for standard bulks; colour-water method for jars — no trace outside the seal area
Decoration adhesion3M 600 tape over print/hot-stamp; 3M 810 over cross-hatched spray & plating (2 mm grid)Pressed 10 passes, pulled at 45° — no lift; hot-stamp single-point loss ≤0.5 mm²
Thermal cyclingOven and freezer, filled48°C and -15°C, 24-48 h each — no cracking, warping, leakage or colour change after 2 h recovery
Drop testFilled, onto concrete76 cm to 1.5 m by format (bottles with closure: 1.5 m at 45°; compacts: 76 cm ×3) — no functional damage
Pump & sprayer lifeContinuous actuation, dosage weighed over 50 strokes ×3≥1,500 cycles; priming ≤5-10 strokes; lotion pump output ≥95% (aqueous) / ≥90% (non-aqueous); spray 0.1-0.2 ml per shot, even mist
TorqueTorque meter on closuresOpening 0.2-0.7 N·m (standard caps), no thread strip at rated tightening; application torque matched to your filling line
Formula compatibilityComponent in contact with your actual bulk48°C / 30 days plus ambient & cold controls — no discolouration, odour, swelling, cracking or decoration attack
ColourAgainst signed golden sample, D65 standard lightΔE within the tolerance agreed per material; family components judged side by side in one hand
Heavy metals & hygieneThird-party lab where formats requirePb+Cd+Hg+Cr(VI) ≤100 ppm (EU 94/62/EC); food-contact and REACH documentation on request; microbial counts for wet-contact formats
Transit protectionCarton and pallet spec for exportCorrugated grade specified per weight; glass packed with internal suspension; breakage allowance agreed up front, not argued afterwards

Acceptance is arithmetic, not argument. Defects are classed critical / major / minor with AQL agreed before production (critical effectively zero; sampling to ISO 2859-1 general level II). Surfaces are graded A / B / C — the face a customer sees on shelf is held to a tighter standard than the base — and the defect allowances differ accordingly, in writing.

When Something Fails

Solve first, apportion second.

Our after-sales principle is fixed: the problem gets solved before responsibility gets discussed. The golden sample, the signed specification and the gate records exist precisely so that the second conversation is factual. Systemic issues get a corrective-action loop with the factory (8D-style), and the fix is verified on the following batch — a complaint that only produces an apology has not been closed.

And one honest boundary: we are a sourcing partner, not a single factory. That is exactly why the system above exists — a trading company without published QC gates is asking you to take quality on faith. We would rather you take it on evidence.

Questions

Frequently asked

What is a golden sample in cosmetic packaging?
The physical reference sample you approve and both parties sign before production starts. It fixes appearance, colour, decoration and function; production is checked against it, not against a photograph. Colour tolerance is agreed in writing (ΔE under D65 light), the sample is renewed on an agreed cycle, and a sealed duplicate is retained for reorders. No signed golden sample, no production.
What are the four inspection gates (IQC, IPQC, FQC, OQC)?
IQC checks materials and components before production; IPQC inspects on the line — where compatibility, adhesion and leak problems actually surface; FQC checks finished goods against the golden sample; OQC samples the packed order to agreed AQL before shipment. An order that only gets a final glance has skipped three of the four gates.
Which laboratory tests does cosmetic packaging actually need?
The core battery: vacuum leak (-0.08 MPa, 15 min, loss ≤0.5%), tape adhesion (3M 600 / 3M 810 with cross-hatch, pulled at 45°), thermal cycling (48°C and -15°C), drop (76 cm-1.5 m), pump life (≥1,500 cycles, stable dose), and compatibility with your actual formula (48°C / 30 days). Colour is judged at ΔE under D65 — never by eye under shop light.
How is formula compatibility with packaging tested?
Your actual bulk — not water — sits in contact with the component at 48°C for 30 days beside ambient and cold controls, watched for discolouration, odour, swelling, cracking and decoration attack. The known failure rules: alcohol above ~30% attacks PET (use glass or suitable copolyester), acetone needs glass with a resistant liner, pH below 4 wants glass or PP, oxygen-sensitive actives need high-barrier EVOH structures. Testing catches the exceptions the rules miss.
What does AQL mean on a packaging inspection report?
The Acceptable Quality Limit defines how many defects of each class a statistically drawn sample may contain before the lot fails. Typical cosmetic packaging runs AQL 0 for critical, 1.0-1.5 for major, 2.5-4.0 for minor, sampled to ISO 2859-1 general level II. Agreeing AQL in advance makes acceptance arithmetic instead of argument.
Who is responsible when packaging fails after delivery?
Our rule: solve the problem first, apportion responsibility second. The golden sample, signed spec and inspection records make the second part factual. Systemic issues get an 8D-style corrective loop verified on the next batch. A supplier who opens with whose fault it is — rather than what you need now — is telling you how the relationship will run.
Ready to source?

Get packaging that arrives the way you approved it.

Tell us your format and quantity. Your quote comes with the QC plan attached — gates, tests and thresholds for your specific components.

Reply within 24h · NDA on request · Your details stay confidential

Keep Reading

Related reading.