Standards-based recommendations and QA
Every recommendation we make is tied to a published test method rather than to experience or a rule of thumb. That is the difference between a specification you can defend to an auditor and an opinion you cannot.
The standards, and what each one is for
ASTM D4169 distribution cycle - the sequence of hazards a pack meets in transit ASTM D642 compression resistance of shipping containers ISTA 3A parcel: drops, vibration, atmospheric conditioning ISTA 6 member-specific e-commerce profiles ISTA 7E thermal transport testing for temperature-controlled shipments ISO 2859 acceptance sampling by attributes - AQL, sample size, acceptance number DIN 55468-1 board grades, flute profiles, dimensional tolerances DIN EN 20187 standard test climate: 23 C, 50% RHChoose the profile from the lane
The most common testing mistake is not a bad test - it is a valid test of the wrong journey. A parcel profile tells you nothing useful about a six-week sea lane, and an export cycle is irrelevant to a courier network. Pick the cycle that matches how the pack really travels, including the parts nobody mentions: the port dwell, the restacking, the humid yard.
Conditioning is not optional
Corrugated strength moves with moisture, so a strength figure without a climate is not a figure. Everything is conditioned to 23 C / 50% RH before test so results are comparable to each other and to your supplier's. Then derate to your real environment - because that reference climate is a measurement convention, not a description of your warehouse.
Acceptance: turn opinion into a number
You cannot inspect every box, so lot acceptance runs on ISO 2859 sampling. The essential step is classifying defects by consequence before anything ships:
- Critical - endangers health, breaches a legal requirement, destroys the packed goods or halts a line. No acceptance number; grounds to reject.
- Major - the pack fails its function, or costs real production efficiency. The tight end of the AQL range.
- Minor - a general quality reduction with no significant consequence. The loose end.
With classes and AQLs agreed, "these boxes look bad" becomes "this lot exceeded the acceptance number for its class" - and only one of those can be resolved without a meeting. Draw the sample across the whole delivery, and evaluate only the most serious defect on any one box: you are counting defective units, not defects.
Tolerances and warp
Dimensional tolerance scales with size, measured between creasing centrelines or from a crease to the associated outer edge. Machine-erected cases frequently need tighter tolerances than the general ladder - specify that at quotation. Warp is measurable rather than aesthetic: maximum elevation divided by the distance between supporting edges, expressed as a percentage. Once it is a number, it can have a limit.
What you should be able to ask for
- The test report for the profile matching your lane, with the conditioning stated.
- The sampling plan and the agreed defect classes.
- The specification: grade, flute, construction, coating, tolerances, flute direction.
- A change-notification commitment, so no substitution arrives silently.
If a supplier cannot produce these, the quality conversation will happen later, at the dock, with less information and more money at stake.