Plastic Bottle Sampling: What the Sample Has to Prove, and What Approval Releases
A sample is not a preview. It is the moment the mould, the material, the colour and the print are proved together — and the reference every production batch is held against afterwards.
Approve the sample you can measure, not the sample you like
Sampling from an existing mould typically runs 3–10 days from a confirmed brief. A new tool is a longer programme, because the tool has to be designed, cut and trialled before a sample exists.
What approval does is fix the reference. After approval, the shape, the colour, the finish and the print position are all judged against the sample you signed off — which is what makes a fourth reorder match the first. That is why the round is worth running properly rather than quickly.
This guide covers the sampling decision: what the round proves, what to check, and what approval releases. The tests themselves are on packaging testing, and the inspection that runs in production is on the packaging QC checklist.
Six things a sample round can prove, and they are not all proved at once
A single sample can be perfect on the shape and wrong on the colour. Reading the round as six separate proofs is what stops a partial approval from becoming a full one.
What the round has to prove
The nine checks to run before you approve a sample
Run these in order and write the result down. A sample approved on three of the nine is a sample that will be argued about at the next reorder.
Sample approval checks
- ☐Capacity — fill it to the fill height your line uses and confirm the nominal volume and the headspace.
- ☐Weight — weigh it and compare against the approved specification within ±2%.
- ☐Height and diameter — against the drawing, with the datum stated so both sides measure from the same origin.
- ☐Neck — fit your actual cap, pump, sprayer or dropper, and check the thread engages without cross-threading.
- ☐Torque — open and close repeatedly and check the closure holds its window.
- ☐Leak — fill and invert. A leak on the assembled unit is a reject, regardless of how the bottle behaves on its own.
- ☐Colour and finish — against the physical master sample, in daylight rather than against a screen.
- ☐Print position and registration — against the artwork, measured from the agreed datum.
- ☐Surface defects — scuffing, sink marks, flash and print artefacts, judged against the standard sample rather than against expectation.
How many rounds are normal, and who pays for them
Sampling itself is not free. It covers material, machine setup and courier costs, and pretending otherwise simply moves the charge into the unit price. What matters is agreeing the cost and the number of rounds before the first one starts.
| Round | What it normally covers | Who should pay | What it changes |
|---|---|---|---|
| First round | Proving geometry, material behaviour, colour and decoration for the first time. | The buyer, unless a commercial arrangement says otherwise. | Most of the specification is settled here, or the round shows which line was wrong. |
| Correction round | A single adjustment — colour, a dimension fine-tune or a closure swap. | One free round of adjustments is normal when the correction is inside the agreed brief. | Confirms the adjustment and produces the reference sample. |
| Change round | A fundamental change — a shape redesign, a different material, a new decoration method. | The buyer, because the brief changed rather than the work. | Re-quoted, because it is a new programme rather than a correction of the old one. |
| Approved sample | The reference for the whole production run and every reorder. | Retained by both sides. | Fixes the standard. This is the round that matters. |
- ▸What the sampling charge covers, and whether it is credited back against a first production order.
- ▸What counts as a correction and what counts as a change, because that distinction decides who pays for round two.
How to run the round so it converges instead of looping
Sampling loops are expensive because they are usually self-inflicted. Almost every extra round traces back to a brief that was open when the round started.
Close the brief before the round, not during it
Capacity, material, neck finish, closure and colour decided up front. An item can be open and named as open, but it cannot be assumed.
Say what the sample has to prove
State the checks in advance — capacity, weight, fit, leak, colour, print position. A round with a stated acceptance is a round that ends.
Measure rather than judge
Dimensioned checks against the drawing, weight on a balance, leak on the assembled unit. Photographs are useful for appearance and useless for tolerance.
Send one consolidated response
One message with every observation beats five messages with one each, because a single round can address several items at once.
Approve in writing, and say what is approved
Approve the physical sample against the stated standard, and say explicitly that the production run is held to it.
Keep the approved sample and the colour chip
Archived by both sides, so a question raised a year later is answered by comparing the actual part rather than by reconstructing it from records.
What changes when the sample comes off a new tool
The difference between a stock-mould sample and a first article off a new tool is not the size of the sample. It is what the sample is being used for.
| Sample from an existing mould | T1 first article from a new tool | |
|---|---|---|
| What it is | A production part off a cavity that already exists. | The first shot off a tool built to your drawing, measured rather than judged. |
| Schedule shape | Typically 3–10 days from a confirmed brief. | A longer programme, because the tool is designed, cut and assembled first, and the schedule is confirmed against your geometry rather than assumed from a range. |
| What it can change | Colour, decoration and closure. The geometry is fixed by the cavity. | Geometry as well as colour and decoration, which is why a design change is cheapest before this point. |
| What it fixes | The standard for colour, decoration and assembly on a known geometry. | The reference for the whole run, including every dimension the tool will ever produce. |
| If it is wrong | Usually a colour, print or closure adjustment. | A tool correction and a re-sample, which is why the design review happens before the tool is cut. |
On a new tool, the sample is also the last cheap moment to change anything. After it is approved, the geometry is in steel — and a change after that is a tool modification rather than a decision.
What sampling cannot prove
A sample round is a proof of the specification, not a proof of the programme. Four things sit outside it, and pretending otherwise is how expectations get set wrongly.
- ▸Long-term stability. A sample cannot show how a pack behaves after twelve months on a shelf with the real formula. Stability is your own shelf-life study, and we can supply the samples and the material data for it.
- ▸A full transit simulation. We drop test and leak test, and we test packed units where the route justifies it. A complete transit simulation through a specific route is normally a third-party study.
- ▸Regulatory approval. A sample does not approve a pack for a market. Migration and food-contact testing, labelling rules and registration are separate, and your own regulatory filing is yours.
- ▸A production average. A sample proves what is achievable; the run is what is delivered. That is what the three inspection gates and the QC report per lot exist to cover.
The five sampling mistakes that add a round
- ▸Approving from a photograph. A sample has to be measured and held. Approval sets the reference, and a photograph is not a reference.
- ▸Approving without the closure fitted. The decorative sample can be perfect while the assembly leaks, and approval only covers what was approved.
- ▸Changing the colour after approval. A colour change restarts the round and can move the quantity step, because custom colour is a masterbatch decision.
- ▸Sending four separate messages with one observation each. One consolidated response can be answered in one round; four cannot.
- ▸Leaving an item open and assuming a value. The assumption becomes the tool, and the tool is the expensive place to discover it.
Tell us what the sample has to prove
- ▸The brief, with any open items named rather than assumed.
- ▸What the sample has to prove — the checks you intend to run on it.
- ▸The fill height your line uses, so capacity and headspace can be checked.
- ▸The closure, pump, sprayer or dropper you will fit, or a physical sample of it.
- ▸The colour reference and any artwork, so the decoration proof can be prepared with it.
- ▸How many samples and how many unpainted blanks you need for your own testing.
We normally send 5–10 finished samples plus 5–10 unpainted blanks so you can run your own decoration and compatibility testing alongside ours. If your line needs more, say so and it is quoted with the round.
Sampling, testing and approval reading
Plastic bottle sampling questions
How long does sampling take?+
Sampling from an existing mould typically runs 3–10 days from a confirmed brief. Where new tooling is required, tooling plus a sample run is a longer programme, and the schedule is confirmed against your geometry rather than assumed from a range.
Is sampling free?+
No. Sampling covers material, machine setup and courier costs, and we charge for it rather than hiding the cost in the unit price. What matters is agreeing what the charge covers and whether it is credited back against a first production order.
How many samples will I receive?+
We normally send 5–10 finished samples plus 5–10 unpainted blanks, so you can run your own decoration and compatibility testing alongside ours. Additional samples can be requested at cost.
How many rounds are normal?+
One round to prove the specification, and normally one correction round if a single item needs adjusting. A third round usually means the brief changed rather than the work, and it is re-quoted because it is a new programme.
What should I check before approving a sample?+
Nine things: capacity at your fill height, weight within ±2% of specification, height and diameter against the drawing, neck fit with your actual closure, torque, leak on the assembled unit, colour against the master sample, print position and registration, and surface defects.
What happens if the sample is not right?+
Where the correction is inside the agreed brief, one round of adjustments is normal. Where a fundamental change is needed — a shape redesign, a different material or a new decoration method — it is a change rather than a correction and is re-quoted.
Do you test the sample, or do I?+
Both. We measure capacity, weight, wall thickness at four points per cavity, neck fit against your closure, and run application and removal torque and an air-pressure leak test at 0.6–0.8 MPa with the unit inverted, on the assembled pack. You should test with your own formula and your own line.
What does approval actually release?+
It fixes the reference. Production is run against the approved sample and the colour chip, and every reorder is judged against the same object. That is why the approved sample is retained by both sides rather than consumed in testing.
Tell us what the sample has to prove
Send the brief and the checks you intend to run. We will come back with the sampling schedule, what the round covers and what it costs — before it starts, not after.