A glass bottle trial order is a test run built to remove one named uncertainty before you commit to production, so design it backwards from that uncertainty. Decide what has to be verified, size the quantity as setup units plus a steady run plus retention samples plus test and shipping units, pick the bottle form that can actually prove the point, and agree in advance which records will pass into the production run. A trial planned this way ends in a decision and a file of evidence; one ordered as "a small first order" ends in a tray of bottles.
Two neighbouring stages are easy to mistake for a trial. Drawings, renders and the approval sample belong to the mock-up and design sign-off stage, which comes earlier. Placing and tracking the full order is covered in our outline of the production order process, which comes later. The trial sits between them: the look of the bottle is settled, and what remains open is whether it behaves on your line and with your product.
Start from the uncertainty the trial must remove
A sample tells you whether the design looks right. A trial tells you whether the pack works, again and again, under the conditions the real product will face. Buyers who approve a handsome sample and go straight to production are the ones who learn a full cycle later that the bottle jams at the labeller, the cap weeps after two weeks on the shelf, or the print flakes along the shoulder edge.
So the object of a trial is not the bottles themselves. It is to put the pack through a system: filler, capping head, labelling or sleeving station, carton, pallet, container, warehouse, and then a shelf where the filled product stands for months. Any of those can reject a bottle that looked flawless on screen.
Before asking for a quotation, complete this sentence in your enquiry: "We are running this trial because we do not yet know ____." Everything else follows from it, including the quantity, the bottle form, the tests and the cost, and whether a trial is worth running at all. If the best you can write is "to see the quality", you are asking for a look, not a test.
A trial also has a quieter job that pays off later. It creates the reference items that protect buyer and supplier for the rest of the project: a sealed reference bottle, a record of measured dimensions and an agreed inspection standard. With those in place, a dispute over what was approved is settled by opening the reference. Without them, the same points tend to be reargued at each reorder.
The five things a trial run can verify
Most worthwhile trials test some combination of five dimensions. The quantity each one needs is not a fixed figure. A check that relies on inspecting a few pieces needs a few pieces. A check that relies on how a machine behaves needs enough bottles to set the machine up and then run it steadily, because a jam that shows up once in twenty bottles will never show up in a handful.
| Dimension | What is being tested | Quantity that gives a valid reading | Pass criterion | Carried into production |
|---|---|---|---|---|
| Line fit | Conveying, guide rails, star wheels, filler nozzles, capper head reach, glass-to-glass contact, and any repeat jam at a particular station | A complete line setup followed by a sustained run at target speed | Agreed speed held for the agreed run length with no manual intervention and no repeatable jam | Bottle drawing with critical dimensions; the line settings that ran clean |
| Closure and seal | Whether this cap, liner and neck finish hold a seal, and whether torque or vacuum is still there once the closure has settled | Units sampled across the run, with a subset held back for re-testing at an agreed interval | No leaks by the agreed method; readings inside the agreed window straight after capping and again after settling | Neck finish, cap and liner specifications; the torque or vacuum window |
| Decoration | How screen print, sleeve, frosting or coating sits on real glass and real curvature, and how it copes with filling and handling | Decorated units covering every print position, plus spare pieces for adhesion and abrasion checks | Appearance matches the approved reference; adhesion and abrasion results within the agreed limit | Approved decorated reference; artwork or plate record |
| Packing and transit | Whether divider, carton and pallet pattern protect the glass, and whether the packed load survives the planned route | Cartons for a full pallet layer; a trial shipment if the route or handling is new | Breakage no higher than the agreed allowance; packed weight and case dimensions to specification | Packing specification, carton marking, pallet pattern |
| Filled-product stability | How closure, headspace and decoration interact with the product during the first weeks after filling | Filled units stored under the intended conditions and opened at intervals through the observation window | No unexpected change to product, closure or decoration within that window | Filled reference units and a note of their storage conditions |
Two rows get cut from trial plans more often than they should. Packing and transit is usually assumed instead of tested, yet it causes the costliest failures: a pallet that collapses inside a container takes the entire shipment with it, not a single bottle. Filled-product stability is the one check nobody can speed up, because time on the shelf, not the glass plant, supplies the answer.
Working out the trial quantity
A trial quantity is built up from four parts. Leave one out and a gap appears later.
- Setup allowance. Every filling, capping, labelling or sleeving line uses up bottles before it turns out good ones. They go on setting rail heights, dialling in the capper, registering the print or sleeve and establishing the first-off standard. These are the cost of finding the machine's settings, and if they are not in the quantity the run ends before the line is ready.
- Continuous run. Once the line is in a steady state, the bottles that follow are the only ones that test repeatability, since only here can a drift, a jam or a torque trend develop. Stopping at setup says nothing about the line. A long run with no settings log cannot be reproduced.
- Retention samples. Sealed, labelled units kept as the physical record of what was approved. They resolve arguments over appearance, give a reorder something to be compared against, and can be opened if a filled product misbehaves months on. Take them from the production-intent sequence. Hand-picked good pieces do not represent what the line makes.
- Test and shipping allowance. Destructive tests and transport trials consume bottles, and so can the journey. If some units are also going to a customer, an approval panel or a test house, count those in now.
Put the resulting number in the enquiry together with the reasoning. Asking for "a few hundred bottles" hides the intent and tends to bring back a quotation costed as a sample. A request that sets out setup, run length, retention and test allowances lets a technical team respond precisely, and it brings hidden costs to the surface while they are still a planning matter.
Stock mould, adapted mould or pre-tooling batch
The form the trial bottles take decides what the trial is able to prove. There are three workable choices.
| Form | What it is | What it proves | What stays open | Best used when |
|---|---|---|---|---|
| Stock bottle | An existing stock mould, no new tooling | Line, closure interface, decoration method, packing and transit | The exact final shape, since a stock mould only approximates it | The doubt concerns process: neck size on the filler, finish on the capper, weight on the pallet |
| Revised or adapted | An existing mould with a change, such as another neck finish, a modified height, a different base or changed decoration | The real interface your line will meet | Dimensional approval, which is provisional until the production tool exists | Most serious projects that want a true interface test without a fully new mould |
| Pre-tooling small batch | A limited batch of a new shape made before the production tool is finalised | Filling, sealing, packing and decoration on a representative form | Final confirmation on the production tool | A dimensional or handling risk too serious to find after full tooling |
A stock trial is the quickest of the three and the most economical way to catch a line problem that would otherwise surface after a tool had been cut. The adapted trial is the workhorse; its report must state plainly that dimensions are provisional. The pre-tooling batch tells you the most and costs the most, so before choosing it, settle how its cost is handled if the project goes ahead and how it is handled if it stops.
Name the form in the trial document. "Trial approved" covers a process when the bottles were stock, a provisional interface when they were adapted, and a near-final form still awaiting its production tool when they came from a pre-tooling batch. Teams that let these blur end up arguing about the scope of an approval well after the trial has closed.
How a trial quotation differs from a production quotation
A trial quotation is not a preview of the production unit cost. Production pricing assumes a stable, repeatable run in which tooling, setup, decoration preparation and quality control are shared across a large quantity. A trial carries the same fixed items across a small one, so the cost per bottle is higher by construction. That difference says nothing about how competitive the production figure will be. All figures are quoted per project.
Ask for three lines to be shown separately, because each is treated differently afterwards:
- One-time preparation. Mould preparation or adaptation, decoration plates or screens, artwork proofing, and the engineering time spent defining the trial.
- Trial unit cost. This reflects the small quantity, the manual attention a short run needs and the heavier inspection of a validation run. It is not a production figure scaled down.
- Documentation and testing. Destructive tests, dimensional records and any outside laboratory work. It is sometimes the smallest line and the most useful, since it is what turns bottles into evidence.
Then agree four points in writing before work starts. An email is enough, and it heads off most end-of-trial disputes.
- Whether any credit applies against a later production order, and what triggers it.
- Which parts of the specification are fixed by the trial quotation and which may still change.
- Who keeps the retention samples and the test records.
- Where things stand if the buyer abandons the trial once preparation costs have been incurred.
Mistakes that waste a trial run
- Using it as a negotiating tool. A trial removes technical doubt. If the open question is commercial, it adds cost and delay and settles nothing, and its unit cost makes a weak bargaining reference because it is not a production figure.
- No written protocol. Pass criteria, measurements and storage conditions that are not on paper before the bottles arrive get decided once the results are known, and by then they are no longer criteria. Draft the protocol during quotation and share it with both sides.
- Testing the bottle alone. The bottle is one part of a pack. Examining it empty, without filling, capping, labelling, packing or shipping it, covers a fraction of the risk and gives false comfort.
- Sending only the best pieces. Cherry-picked units conceal the very thing the trial was meant to find. Send bottles from across the run, first-off and last-off included, and say so in the covering note.
- Calling approval final too early. On an adapted or pre-tooling trial, dimensional approval is provisional by definition. Recording it as final commits both sides to something the production mould, cut to its own tolerances, may not match.
Deciding whether to go ahead
Judge the results with three tests, taken in this order.
Reproducibility. Did the passing attributes come from the process or from selection? If an operator picked out the good bottles while the run as a whole had jams, leaks or print drift, the trial was survived, not passed. Ask what proportion of the run met the requirement and whether that came from a stable setting.
Root-cause clarity. Each failure needs an identified cause and a correction that one of the two parties controls. A leak traced to the wrong liner is solved. "Some units leaked" is not. A failed trial with a clear cause can move forward; a passed trial with unexplained defects cannot, because the next run inherits the same hidden instability.
Commercial fit. Check that the validated configuration still meets the original commercial brief. A trial can work technically and fail here, for example when the sealing fix calls for a closure whose supply position is longer than the launch schedule permits, or when a packing change takes the delivered case weight over a distribution limit. This test comes last, and it is not optional.
Repeat a trial that failed on reproducibility or root cause only once something has changed, such as a different liner, a corrected rail setting or a revised mould. An unchanged trial returns an unchanged result at full cost. A failure on commercial fit usually calls for a design change, which sends the project back to the drawing stage instead of to another sample.
The handover package for production
Alongside the decision, a successful trial has to hand over information. Otherwise production begins from a conversation and each parameter is found all over again. The minimum set fits on a single page, and it is worth listing before the trial starts so the right readings are taken during it.
| Item | What it contains | Why it matters |
|---|---|---|
| Sealed reference sample | Units sealed and labelled with trial reference, date and run identification, held by both parties | It is the physical definition of the approved appearance and the first thing checked when a later batch is queried |
| Dimension record | Measured critical dimensions with instrument and method stated | A drawing shows the intent; the record shows what the approved run was, which gives an inspection limit its meaning |
| Decoration reference and artwork record | The approved decorated unit and the version of artwork, plate or screen behind it | Decoration drift is hard to spot from a description and easy to spot against a sealed sample |
| Closure specification and sealing window | Cap and liner specification, the torque or vacuum values shown to hold, and the conditions of measurement | A torque figure with no conditioning interval or instrument is not a specification |
| Packing specification | Divider type, carton dimensions, pallet pattern, layer count and any stacking limit confirmed in the transit trial | It fixes the pack that was actually proven on the route |
| Inspection standard and acceptance limits | Attributes inspected, inspection intensity and limits, set out for instance as an AQL plan with critical, major and minor defects defined | Defect classes are stated instead of implied |
| Trial report | A short account of what was tested, observed, failed, changed and concluded | It explains to a future colleague why a parameter was chosen, so a reorder need not repeat the exercise |
After the trial, the questions turn commercial
A trial that has done its job raises the next set of questions: what quantity, on what terms, against what delivery pattern, and from how many sources. The technical file then becomes an attachment to the discussion, which shifts to quantities, packing, terms and scheduling.
If your programme has grown beyond a single project, supply structure matters more than trial design, and the range of empty glass bottles for wholesale programmes is the place to continue. If your schedule hinges on how long the validated configuration takes to build, see our notes on planning the interval from confirmed specification to delivery. A sealed reference, a dimension record, an inspection standard and a written report already cover most of what those stages ask for.
When you brief us on a trial, two things let us propose a design and acceptance items against your real conditions: the verification goal, phrased as the uncertainty to be removed, and a short description of your filling and packing line, covering the container size range it handles, the capping or sealing method and any known handling constraints.
Frequently asked questions
How many bottles should a trial order include?
Add up four needs: bottles used in line setup, bottles in the sustained run, sealed retention units, and bottles consumed by destructive tests and shipping. The run has to be long enough for a recurring fault to show, since an interference occurring once in twenty bottles will not appear in five. Show the sum in your enquiry so the quotation matches the purpose.
Can the trial use stock bottles instead of a new mould?
Yes, and on many projects that is the better choice. Stock bottles answer the process questions behind the most expensive failures, namely line handling, sealing, decoration and transit, with no new tooling. They cannot confirm the final shape, so dimensions stay open until the production mould exists. For a truly new design, a stock trial followed by a pre-tooling batch is a common and efficient sequence.
Is the trial cost credited against the production order?
It is a commercial term that differs from project to project and should be agreed in writing beforehand. Keep preparation, trial unit cost and testing apart. Preparation is often the creditable part, while the unit cost pays for work already done and usually is not. Record which part is creditable, on what conditions, and what applies if the project does not continue.
Why do trials most often fail to be useful?
The leading cause is a missing protocol: with criteria, measurement method and storage conditions left undefined, results are interpreted after the fact and the trial turns into an exchange of opinions. The runner-up is submitting only the best bottles from the run.
How long does a trial order take?
Timing is set per project, and the form of the trial drives it far more than the quantity does, because stock bottles and a pre-tooling batch go through entirely different preparation. Ask for the schedule split into preparation, production and any testing or transit stage, so you can see which part you can influence.
What must be signed off before production starts?
Everything in the handover package: sealed reference, measured dimensions, decoration reference with artwork version, closure specification with sealing window, packing specification, inspection standard with acceptance limits, and the trial report.
Is a trial order just a larger sample?
No. A sample shows appearance. A trial shows behaviour on the line, at the seal, through packing and transit, and after filling. Companies that use one word for both tend to order a sample when they needed a trial.