Lot versus batch: clearing up the terms

Batch and lot are used as if they were synonyms, and most of the time nobody pays for the confusion. The cost arrives with the paperwork: on a certificate of analysis, on a label, in a receiving record, where one of the two words is doing work the other cannot. The distinction is short, and once it is made, the documents that use the terms start reading differently.

A batch is an event

A batch is a specific act of manufacture or preparation. It has defined inputs, a defined process, a defined time, and it is finished when the run is finished. Two batches of the same material, made to the same procedure on different days, are similar; they are not the same population. The batch is what a batch record documents — what went in, what was done, what was produced — and the record is specific to the event by construction.

A lot is a designation

A lot is a quantity of material identified as uniform for the purpose at hand. It may be one whole batch, part of a batch, or several batches blended and mixed so they can be treated as one. Unlike a batch, a lot is not something that happens; it is a decision someone makes and documents — here is the boundary, here is the population that will be sampled and sold under this number. That decision is what the sampling plan describes, and it is why the plan matters.

Where the terms drift

Some industries print batch and lot on the label as a single code, because for that material they coincide. Some suppliers use “lot number” for what a manufacturer calls a batch, and add a second number for the repacked quantity. None of this is wrong. What is wrong is ambiguity: a document that says “batch” while meaning the blended, divided, repacked population has left the reader unable to know what the results describe. A short glossary — two lines, in the laboratory’s own procedure — settles it for the site, whatever the supplier’s habits.

What the certificate attaches to

A certificate of analysis attaches to a lot, because the lot is the population the sample was drawn from. Read which one it names. If the lot came from a single batch, the certificate is a statement about that batch and nothing more is needed. If the lot is a blend of several batches, the certificate should say so or the sampling records should exist somewhere, because a composite sample tests the blend, and the blend is the lot. A result that is fine for a well-made blend can be misleading for one batch inside it, and the reverse.

Reading them together

The practical checks follow from the definitions. The lot on the certificate, the lot on the label and the lot on the paperwork should match character for character. Where a container was split into sub-lots or repacked, the connection between the old lot and the new one should be recorded, because splitting ends the sampling assumption. Where a certificate names a batch for a blended lot, ask how the blend was sampled, and let the answer be written rather than remembered.

The words are cheap; the populations are the point. When a reader knows what population a result is meant to represent, batch and lot stop being a vocabulary problem and become what they should have been all along: two precise answers to two different questions.

Repacking deserves special mention, because it is where good intentions produce ambiguity. When a lot is divided, every new container should carry a number that leads back to the parent lot, and the division should be dated. If the material is blended or subdivided again, the chain of designations grows — and each link that is written down preserves the answer to the only question that matters later: what population did the test describe.