A 4 g difference on a 30 g bottle is usually not cheating — it is a tolerance that nobody put in writing. The grams did not disappear into someone’s pocket; they moved because blow molding and injection molding always produce a spread. The real problem is not the missing weight. It is that “30 g” was treated as a promise when it was only ever a sample reading.
Why a bottle weight is a range, not a number
When a quotation says 30 g, that figure comes from one bottle, pulled from one cavity, at one moment in a production run. It is a snapshot, not a guarantee. Every process that shapes plastic has natural variation: melt temperature drifts a little across a shift, the parison sags slightly differently as the die heats, and each cavity in a multi-cavity tool fills a fraction differently from its neighbours. On a 2-cavity tool the two cavities can easily sit 1–2 g apart on the same run, and both will be considered normal on the floor.
This is why experienced buyers never accept a single number. They accept a target plus a tolerance — and they write both down. Without the tolerance, you and your supplier are each holding a different definition of “correct”, and you will only discover that when the cartons are already on a boat.
Where the missing grams actually go
Before assuming bad faith, it helps to know the four legitimate places weight comes from.
Cavity-to-cavity spread. In a multi-cavity tool the outer cavities often run marginally lighter than the centre ones. On a well-maintained tool the spread is small; on a tired tool it widens.
Material distribution. In extrusion blow molding, wall thickness is controlled by parison programming and by where the bottle needs strength. A bottle with a thin shoulder and a thick base weighs differently from one programmed the other way, even at identical settings.
Regrind ratio. Off-spec bottles and the flash trimmed off a blow-molded part are commonly reground and fed back in at a controlled percentage. Regrind has slightly different flow behaviour, which nudges weight. A controlled regrind percentage is normal industry practice; an undisclosed one is not, and it belongs in your specification.
Process drift across a long run. A run that starts at 30.2 g can settle at 28.6 g by the end of a shift if nobody is checking. This is the most common cause of a whole shipment arriving light, and it is entirely preventable with routine weighing.
What a fair tolerance looks like
There is no universal standard that applies to every cosmetic bottle, and anyone who quotes you a hard number without knowing your process is guessing. What is widely seen in the industry as a workable starting point is a tolerance of roughly ±5% on a bottle in this size range, with tighter bands agreed when the bottle is very light or the closure fit is sensitive. For a 30 g bottle that is roughly 28.5–31.5 g as a common industry band — treat it as a discussion opener with your own supplier, not as a rule, because the achievable band depends on the tool, the resin and the machine.
The important part is not the exact percentage. It is that the band is agreed before production, written into the specification sheet, and measured the same way by both sides.
Why underweight hurts you more than overweight
Buyers instinctively dislike paying for a light bottle, but the bigger commercial risk is functional, not financial. Gram weight is a proxy for wall thickness, and wall thickness is what decides whether a bottle survives filling, capping, palletising and a warehouse summer.
A bottle that is consistently 10% light may still look fine empty, then panel or dent under hot-fill conditions, or creep at the base when stacked. It can also change how the closure seats, because the neck land and thread are part of the same shot — a thin neck is where cap-retention and leak complaints start. This is why the weight question is worth asking properly rather than waving through.
How to weigh and sample so the result means something
Most weight arguments are actually sampling arguments. A useful check is simple:
Take a random sample across the shipment — pull from several cartons and several pallet layers, not from the top of one box. Twenty to thirty pieces is a practical size for a first check. Remove caps, pumps and any inserts, because a closure can weigh more than the difference you are arguing about. Use a scale that reads to 0.01 g; a kitchen scale will not resolve what you need. Then look at the spread, not just the average — a sample averaging 29 g but ranging 26–32 g tells you the process is unstable, which is a more serious finding than a uniform 29 g.
Record the lot number and the cavity identification if the bottle has one. Cavity marks turn a dispute into a diagnosis: if all the light bottles come from one cavity, you have a maintenance issue, not a cheating supplier.
What to write into the specification
The clause that prevents this argument is short. Specify the target gram weight, the tolerance band, how it will be measured (sample size, caps on or off), when it will be measured (during production, at inspection, or on arrival), and what happens if it falls outside — rework, credit, or rejection. Also state whether regrind is permitted and at what percentage, and require that the supplier keep weighing records for the run.
On our own lines we weigh by cavity at set intervals and keep the record against the work order, so if a customer asks why a lot ran light we can answer with data rather than an opinion. That practice is worth asking for from any supplier, because a factory that cannot produce a weighing record is a factory that was not watching.
When it genuinely is a problem
Weight drift inside an agreed band is normal. Three patterns are not: a whole shipment sitting below the agreed floor; a bimod sample suggesting two different production standards were mixed; and a light bottle paired with higher-than-expected breakage or cap-retention failures. In those cases the weight is a symptom, and the fix is a specification conversation plus a process correction — not an accusation.
We have been making bottles since 2003, and in most of these conversations the honest answer is that nobody defined the number tightly enough at the start. Once the band is in writing, the same supplier usually hits it without difficulty.
Questions buyers ask us most
Is a 26 g bottle when I paid for 30 g grounds for rejecting the shipment?
Only if your written specification said so. Without an agreed tolerance, a 13% shortfall is a negotiation rather than a breach. With a tolerance clause — say a ±5% band — the same shipment is clearly out of spec and you have a defined remedy. This is the single cheapest clause to add to a specification sheet.
Can a factory deliberately make bottles lighter to save resin?
It happens, but it is far less common than drift. Resin is a real cost, so a supplier under pressure on price may let weight slide to the bottom of the achievable band. That is a reason to specify the band, not a reason to assume fraud. Look for a downward trend across lots rather than one light carton.
Does a lighter bottle always mean a weaker bottle?
Not automatically, but weight and wall thickness are closely linked in blow molding, and wall thickness drives top-load strength and panel resistance. A uniform 5% reduction is usually tolerable; a 10% reduction concentrated in the shoulder or base is where dents, creep and closure problems begin.
Should I weigh bottles with the cap on or off?
Off, and say so in the specification. Closures vary in weight and can easily exceed the difference you are measuring, which makes the two sides’ numbers incomparable. Weigh the bottle body alone and record whether any insert or liner was included.
How many bottles should I sample?
For a first check, 20–30 pieces drawn randomly across cartons and pallet layers is practical and will show both the average and the spread. If the spread looks wide, increase the sample before drawing conclusions — a wide spread points to process instability, which matters more than the average itself.
Related reading
If this is your first order from a Chinese bottle factory, these are the pieces buyers read next.
- What payment terms a small brand should expect from a Chinese bottle factory
- How to compare packaging quotes on the same basis
- Why cosmetic bottles leak, crack or craze, and who is responsible
- Cosmetic bottle lead time: where the weeks actually go
- Launching a first skincare line: the packaging roadmap
- Resin Quality Checklist: COA, Migration Testing and EU 10/2011
- Bottle Quality Control: What to Inspect Before Bulk Production
- 7 Red Flags When Sourcing Cosmetic Bottles from China
- Cosmetic Bottle MOQ, Tooling Cost and Lead Time
- How to Choose a Cosmetic Plastic Bottle Manufacturer in China
- The Cosmetic Packaging Supplier Audit Checklist