Zero defects is not a specification — it is a wish. Every production run of plastic bottles contains some level of imperfection, and the useful question is never “are there any defects” but “how many, of what severity, and did we agree the limit before production started”. If you did not agree a limit in writing, you are not inspecting against a standard; you are negotiating after the fact.
Why “make them all perfect” is not an instruction
Buyers often ask for zero defects and suppliers nod, because nobody wants to be the person who argues for defects. The problem is that “zero” cannot be verified without inspecting every piece, which is slow, expensive, and — on a decorative or structural judgement like a faint flow line — still subjective. What actually works in practice is a sampling plan: agree how many units get checked, and how many failures are tolerated before the lot is rejected.
This is the idea behind AQL, the Acceptable Quality Limit used widely in consumer goods inspection. It does not mean “we accept bad bottles”. It means both sides agreed in advance what the pass/fail line is.
What AQL means, and the part everyone gets wrong
AQL is expressed as a percentage or a number, and the most common misunderstanding is reading it as “the defect rate we expect in the shipment”. It is not. AQL is a threshold used with a sampling table: if the number of defects found in the sample is at or below the acceptance number, the lot passes; above it, the lot fails.
A second misunderstanding is assuming a tighter AQL is always better. It is not free. Tighter limits mean larger samples, more inspection hours, more rejected lots that were probably fine, and — because the supplier prices that risk in — a higher unit price. The right AQL depends on how bad the defect would be for you, which is why defects get classified before they get counted.
Classify defects before you argue about them
The single most effective thing a buyer can do is sort defects into three buckets before the inspection starts.
Critical defects make the bottle unusable or unsafe: a cracked neck, a sharp flash edge, a closure that will not seal, contamination inside the bottle. Most buyers treat critical defects as unacceptable at any count, and that is a reasonable position to hold.
Major defects are visible or functional problems a customer would notice and might return: a black spot on the front panel, a visible sink mark, a distorted thread that still caps but looks wrong.
Minor defects are cosmetic and unlikely to be noticed in use: a faint flow line on the base, a barely visible gate mark, slight colour variation within an agreed band.
Separating them changes the whole conversation. A 2% rate of minor marks is usually a non-event. A 2% rate of neck cracks is a rejection. Without the classification, both problems get discussed as “2% defective” and nothing gets resolved.
What levels are typically used for cosmetic bottles
There is no single correct number, and the right level depends on the decoration and the market. As a general industry pattern rather than a rule, critical defects are usually held at zero tolerance, major defects commonly sit in the region of AQL 1.0 to 2.5, and minor defects around 2.5 to 4.0 — with the tighter end chosen for premium or highly decorated packaging where appearance sells the product. Treat these as a starting point for discussion with your supplier and your inspector, not as a specification, because your risk profile and price point should decide it.
The sample size matters as much as the level. A plan that inspects 20 pieces out of 50,000 tells you very little, and both sides will know it. Ask what sample size the AQL level implies for your order quantity, and make sure that is actually what gets inspected.
How to sample so the result means something
Random means random. Pull cartons from different pallets and different layers, not the first carton off the line and not the one the supplier set aside for you. If the shipment spans more than one production day or more than one cavity set, sample across those too, because defects often cluster by time or by cavity rather than spreading evenly.
Inspect under consistent conditions: the same distance, the same light, and the same viewing time per piece. Many cosmetic defects are only visible under good light at arm’s length, which is roughly how your customer will see the product — so that is the standard to use, not a bench lamp at ten centimetres.
What to agree before production, not after
Put four things in writing: the defect classification with concrete examples for your bottle; the AQL level for each class; the sample size and how cartons will be selected; and the consequence of a failed lot — re-sort at the supplier’s cost, credit, replacement, or rejection.
Also agree who inspects and when. A pre-shipment inspection at the factory is far cheaper than discovering the problem after the goods land, because the remedy options shrink dramatically once the bottles are in your warehouse.
On our side, bottles are checked against the agreed classification before shipment, and the inspection record goes with the lot. If a customer finds something different at their end, having both records makes it a comparison rather than a dispute — and in our experience since 2003, most disagreements turn out to be two people using different definitions of the same word.
Questions buyers ask us most
Can I demand zero defects?
You can ask for it, but it is difficult to verify and most suppliers will price the risk into your unit cost. A more practical approach is zero tolerance for critical defects and a defined AQL for major and minor ones, with the classification written down and agreed before production.
What AQL should I use for cosmetic bottles?
It depends on how the defect would affect your customer. A common industry starting point is zero for critical defects, roughly 1.0–2.5 for major, and around 2.5–4.0 for minor, tightened for premium decorated packaging. Confirm the sample size the level implies, because a tight level with a tiny sample does not protect you.
Are black spots a major or a minor defect?
Location decides. A black speck on the front panel where your label does not cover it is normally major, because the customer sees it. The same speck under a full-wrap label or on the base may be minor. Decide these examples before inspection so the inspector is not improvising.
Who pays if the lot fails inspection?
That should be settled in the purchase terms before production — typically the supplier re-sorts or replaces at their cost when the lot fails an agreed AQL. Without a prior agreement, cost allocation becomes a negotiation you are having at the worst possible moment, when you need the bottles.
Is a pre-shipment inspection worth it on a small order?
Usually yes for a first order with a new supplier, because it establishes how that factory’s output actually looks and gives you a baseline. Once a supplier has a record of passing, many buyers move to periodic or risk-based inspection instead of inspecting every lot.