sampling plan
A sampling plan is the written game plan for how to collect samples — decided in advance, on paper, before anyone touches the material. It's like a recipe's shopping list, written before you go to the store: where to go, what to grab, how much, and how many trips, so nothing is left to a hurried guess in the moment.
Formally, a sampling plan specifies how many samples to take, from where and when, how large each should be, by what method and tools, and how they will be handled and stored afterward. It turns the vague aim of "go get a representative sample" into concrete, repeatable instructions that anyone could follow and get a comparable result.
It matters because sampling decisions made on the fly tend to be biased and inconsistent, and the errors they cause cannot be fixed later in the lab. A good plan is built to control sampling error to a target level and to satisfy legal or regulatory requirements; the honest caveat is that designing it well demands knowing how variable the material is beforehand — which is itself sometimes only learned by sampling.
A food-safety sampling plan for a 500-bag flour lot might require drawing 50 g from each of 15 randomly chosen bags, combining them, mixing, and sub-sampling 200 g for the lab — all written down before sampling begins.
Decide where, when, and how many before touching the material.
A sampling plan is the design; sampling is carrying it out. The plan also fixes the number of increments — and statistically, taking more, smaller increments and combining them usually gives a more representative composite than one large grab, because random patchiness averages out.