Deliveries get checked in one of two ways: nothing is checked, or everything is checked. Both are wrong for most packaging. Nothing at all lets bad batches through, and full inspection costs more than the value of the defects found, while producing inspection fatigue that misses things anyway.
Size the plan to the consequence
The right sample size depends on what happens if a defective unit gets through. If a bad unit means a line stoppage, sample aggressively. If it means a crate that is retired next week anyway, sample lightly, or not at all. There is no universal number, and pretending otherwise is why most plans get ignored.
Sample across the delivery, not from the top
Units from the top of the first pallet are the least representative ones available. Manufacturing drift happens over a run, so sample from early, middle and late pallets. Where a delivery has several production batches, sample each – a mixed delivery with one bad batch is otherwise invisible.
Specify what counts as a defect before checking
An inspector with no definition finds defects inconsistently. Write the defects down – as we did for grading returned sheet and pads – ideally with photographs, and the pass or fail decision becomes repeatable between people and shifts.
Decide what happens on a fail
A plan without a defined consequence is decoration. State in advance whether a fail means reject the delivery, inspect one hundred per cent, or accept with a credit note – and who decides. Sorting this out at the time of a failed delivery is how relationships sour.
Record and use the results
Sampling data is only useful if it goes somewhere. A simple log of pass or fail by supplier and part builds a picture within a year that supports both complaints and price negotiations – and it is the evidence that makes a conversation with a supplier short rather than argumentative.
Trust, but re-check after changes
A supplier with a long clean record justifies lighter sampling; that is what the history is for. But any change in material, tooling or sub-supplier resets the history. Ask to be told when those change, and re-sample heavily for the first delivery afterwards.
Sampling is not the same as inspecting for fit
A delivery can pass sampling and still not assemble with the previous one, because sampling checks units in isolation. Where packs interlock, stack or nest, test them together – take a sample from this delivery and a sample from the last and confirm they still work as a set.
This catches the drift that individual inspection misses: a dimension within tolerance on both batches that nonetheless stops them nesting. It takes thirty seconds and it prevents the worst kind of problem, which is discovering incompatibility at the point of use.
Tell the supplier the plan
A sampling plan works better when the supplier knows it exists. Stating sample size and defect definitions in the order turns inspection from an argument into a shared routine, and it quietly improves incoming quality before any inspection happens.
What we can do
We supply PP cable reels and other items against agreed acceptance criteria, and we will tell you when a production change means you should re-check the first delivery. Send the criteria you already apply – or ask us to suggest a plan sized to your consequence of failure.