One cracked bottle in fifty is a rate, not an event, and a rate is decided by a test rather than by an argument. The evidence that settles it is a filled drop test on the pouch exactly as it was shipped, run to a named procedure, with the cartons sampled by position and the first crack recorded drop by drop. Fracture patterns and carton condition tell you where to look. They do not tell you whether a fabric drawstring pouch was ever built to protect a glass bottle, and that is the question underneath the dispute. Run the test and the argument turns into a document: a configuration, a schedule, a result, and a specific ask. Skip it and the same conversation comes back with the next shipment, defended by whoever happens to be holding the file that season. This page is the order to run that test in, the evidence it produces, and how the result becomes a claim or a specification change.
What the Two Sides Are Actually Arguing About
The factory says transit. You say the packing record shows the pouches left in good condition. Both statements can be true, because neither one addresses the object in the middle: a fabric bag with no cushioning, no divider, and a bottle resting against a seam and a base. A drawstring pouch is presentation packaging by default. It is built to be seen, tied and reopened, not to absorb an impact.
That reframing matters because it changes who owns the failure. If the purchase specification never said the pouch had to protect the bottle, the factory did not fail to meet a requirement, and the transit explanation is not a deflection. If the specification did say it, or if the gift set was sold as a shippable unit, then protection is a requirement and the only way to know whether it was met is to test the packed unit. Either way, the argument is about the specification, and the test is how you find out what the specification is actually worth.
The Reproduction Test, Step by Step
The test is only worth running if it reproduces the shipment rather than an idealised version of it. Work through these steps in order, because a test that skips one of them produces a result the other side can dismiss.
- Sample from the delivered batch, not from new stock. Take enough pouches and bottles to see a mechanism, and include units from more than one carton.
- Reproduce the shipped configuration exactly: the same inner polybag, the same number of sets per carton, the same divider or the same absence of one, and the bottle seated the same way round.
- Condition the filled cartons as the chosen procedure requires before the drops. Temperature and humidity change how a fabric, a board and a glass bottle behave together, and a test run straight out of a cold warehouse is not the same test.
- Drop to a named procedure rather than to a height someone remembers. ISO 4180 sets out the general rules for building a performance test schedule for complete, filled transport packages, including which drop orientations belong in the schedule and how the drop height follows from the package mass.
- Inspect after every drop, not at the end. The drop that produces the first crack is the most useful data point in the whole exercise, and a single inspection at the end throws it away.
- Run a control. The same bottles, the same cartons, one change to the protection, tested on the same schedule. Without the control you learn that something broke; with it you learn what fixed it.
What to Record While It Runs
The report is only as good as the notes taken during the test. Two records carry most of the weight, and both are cheap to capture if you plan for them before the first carton is opened.
The as-shipped configuration, photographed before the first drop
Photograph one open carton with the pouches in place and one pouch with the bottle seated, on a plain surface, with a rule in frame. Record how many sets sit in the carton, which way the bottle faces, whether the neck or the base sits at the pouch base, and whether anything wraps the bottle. This photograph is what stops the later claim that the test used a different packing method, and it is the piece of evidence most often missing when a dispute escalates.
Which drop produced the first crack
For every unit that fails, record the drop number, the orientation of the drop, and where the unit sat in the carton: corner, edge, middle, top layer, bottom layer. That last field turns a pile of broken glass into a map. A failure that clusters in one position is a stacking or compression story; a failure spread evenly across positions points at the interaction between the bottle and the pouch itself. The same position logic applies when you build an inspection checklist for fabric pouches, because where a defect appears is usually more informative than how many defects there are.
The Acceptance List and the Evidence Pack
Six pieces of evidence are worth assembling, and each one answers a different question. Knowing which question each piece answers is what keeps a claim from turning into a debate about who is more credible.
| Evidence | What it can prove | What it cannot prove |
|---|---|---|
| Filled drop test on the as-shipped configuration | Whether this packaging survives the schedule, and which drop breaks it | Which individual unit failed in the field |
| Fracture pattern of the returned bottles | Whether a break started from an impact or from a flaw in the glass | How the force travelled through the pouch before it reached the bottle |
| Carton condition photographed on arrival | Whether the outer package took a visible impact | Whether that impact happened before or after the carton was packed |
| Retained sealed samples from each batch | What the pouches were when they left the factory | How the co-packer actually packed them |
| Timestamped packing photographs | What the packing configuration was on the line | That every carton matched that configuration |
| The failure rate split by carton position and pallet layer | Where to look next, and whether the cause is stacking or unit-level | Root cause on its own |
Two of these are usually missing. The retained sealed samples are missing because nobody keeps pouches from each batch, and the timestamped packing photographs are missing because the packing line is not treated as a place where evidence is created. Both are cheap to add and both become impossible to reconstruct after the fact.
Where the One in Fifty Concentrates
A two percent failure rate is small enough that a twenty-unit test cannot measure it. What the test can do is find the mechanism, and the field data can then be used to see whether the mechanism explains the rate. That is why the stratification matters more than the raw count.
If the failures arrive from one corner of one pallet layer, look at stacking and compression before you look at the pouch, because the units at the bottom of a stack carry a load the units at the top never see. If the failures are spread evenly across positions and cartons, the constant is the bottle inside the pouch, and the pouch construction becomes the suspect. If they cluster in one production batch, the suspect is that batch’s fabric or its packing, and the batch code is what lets you isolate it.
What It Costs You to Run
The cost of the test is mostly time and samples: pouches and bottles from the delivered batch, a drop rig or an outside lab, an inspector for the day, and a report. Set that against the cost of a repeat claim, a retailer chargeback, and a second season of replacing sets, and the test is usually the cheaper of the two routes, even before it produces anything you can put in front of a customer.
It is also the only route that produces a document the factory can act on. A complaint that says “the bottles arrive cracked” invites an explanation. A report that says which drop at which height and which orientation broke the first bottle in the bottom corner position invites a change.
When the Pouch Has to Change
If the test shows the as-shipped pouch fails, the options are structural rather than cosmetic. A padded base or a base board stiffens the floor of the bag so the bottle’s own weight is not concentrated on one seam. A divider or a sleeve stops the bottle moving inside the pouch between packing and unboxing. A retention band or a snugger fit removes the travel that turns a small impact into a strike. Moving the seam away from the bottle’s shoulder removes a hard line pressing into glass.
Each variant is tested on the same schedule as the original, otherwise the comparison is meaningless. The construction choices themselves are covered in interlining, stiffeners and base boards for fabric pouches, and the moisture side of the same shipment is handled in moisture, mould and container rain, because a carton that takes on water during the same voyage will fail differently again.
What Goes into the Claim Letter
The letter carries the report, the batch codes, the stratification table, and one specific ask. State the mechanism the test found, state which configuration failed, and state what you want: replacement of the affected units, a credit, a change to the packing method, or a change to the pouch specification. Send the same package to the factory and to your own quality lead, so the version of events that reaches the next order is the one in the report.
If the conclusion is that the pouch was never a protective package, say so in the letter and move the fix to the specification. That is not a concession; it is the finding, and it is the one that prevents the same one in fifty from coming back next season. Any change that follows goes through the same route as change control on an approved pouch specification, because a protective insert added on the packing line and never written into the spec will disappear the first time the line is busy.
Заключение
The dispute ends when the question changes from who is to blame to what the packed unit does under a repeatable drop. Test the pouch as it shipped, record the configuration and the first crack, stratify the field failures by position, and decide from the result whether the pouch is presentation or protection. Then write that decision into the specification, where it survives the next shipment.
Run the test once and you own a document. Skip it and you own a conversation that repeats every season, with a different person defending the same position.
