Batch traceability: codes, labels and custody

What a batch actually is

A batch (or lot) is a quantity of material produced in a single defined operation, under conditions that are assumed to be uniform throughout. The idea is what makes testing economically possible: you cannot test every microgram, so you define a quantity that should be homogeneous, sample it, and let the result stand for the whole.

That inference is only as strong as the definition. If “batch” means one synthesis and one purification, pooled and mixed before filling, a result from one vial reasonably describes the rest. If it means several production runs combined for convenience, or repackaged material of mixed origin, the same result describes far less than it appears to.

Reading batch codes

There is no universal format for a batch code, and no way to decode one with certainty from the outside. Common patterns include a date component (240815 or 20240815), a compound abbreviation, a sequence number for the run, and occasionally a site or line identifier. Some producers deliberately use opaque codes so that production volumes cannot be inferred.

What you can assess is internal consistency rather than meaning:

• Does the code follow the same structure across items you have seen from the same source, or does the format change arbitrarily?
• Does an embedded date, if present, sit in a plausible relationship to the analysis and report dates?
• Is the code unique to the batch, or does the same code appear across products and time periods?
• Is it printed as part of the label or added by hand — and if hand-written, is it legible and unambiguous?

Matching a label to a report

This is the single most useful check in the whole subject, and it takes about a minute. Compare the label and the analytical report field by field:

If any of these fail, the honest conclusion is not that the material is misrepresented — it is that the report does not demonstrably describe it. That is a different and more precise statement than “it has a COA”.

Chain of custody

Chain of custody is the documented history of a sample: who held it, when, and what happened to it at each step. In accredited testing it is a formal requirement, because a result is only defensible if the laboratory can show that the sample it measured is the sample it was given, unchanged.

A complete custody record answers:

• Who sampled the batch, and how? Sampling that is not representative invalidates the inference from sample to batch, no matter how good the analysis is.
• When was the sample taken, relative to production?
• How was it identified and sealed? Unique sample identifiers and tamper-evident packaging exist so that substitution can be ruled out.
• How was it transported and stored? Conditions in transit are part of the result, particularly for materials that degrade with moisture or heat.
• Who received it, and when? The laboratory’s receipt record closes the loop.
• What happened to the remainder? Retained samples allow a disputed result to be re-examined later.

For third-party testing there is an additional and often decisive question: who chose the sample? A sample selected and submitted by the party making the claim is a weaker form of independence than one drawn by the laboratory or by an independent sampler, even when the analysis itself is impeccable. Nothing about that is necessarily improper — it is simply a limit on what the result can support.

Verification records

Behind a well-documented batch there is normally a set of retained records: the production record, the sampling record, the analytical raw data, the issued report, and a register linking the batch code to all of them. Good laboratory practice frameworks exist largely to require that these records be attributable, contemporaneous and retained for a defined period, precisely so that a claim made today can be re-examined later.

As an outside reader you will rarely see the underlying records. What you can observe is whether the system behaves as though they exist:

• Are reports batch-specific, or is one document reused across many batches?
• Do report numbers and sample identifiers appear to be issued sequentially by an organisation that keeps them?
• Can a specific batch code be queried, and does the answer come back specific to that code — or identical for anything you type?
• Are historical reports still retrievable, or does documentation only exist for current stock?

Lookup portals and QR codes

Batch lookup pages and QR codes on labels are genuinely useful: they shorten the distance between a label and a document, and they make substituted or expired paperwork easier to spot. It is worth being precise about what they establish.

A lookup portal demonstrates that the party operating the portal holds a record associating that code with that document. That is a real and useful fact. It is not independent verification: the portal, the record and the claim usually originate with the same party, and a QR code printed on a label can point anywhere. The strongest form is a report hosted by, or independently confirmable with, the issuing laboratory itself.

What a broken chain actually means

Precision matters here, because a broken chain is regularly reported as either nothing at all or as proof of fraud, and it is neither. Three distinct situations:

• Unverifiable. Documents exist but cannot be connected to the material — for example a report with no batch code. Nothing has been shown to be wrong; nothing has been shown at all.
• Inconsistent. Documents contradict each other — a report predating its batch, or two different results for one code. Something is definitely wrong with the documentation, though not necessarily with the material.
• Contradicted. An independent measurement disagrees materially with the certificate. This is the only case that says something direct about the material itself, and it requires an independent measurement to reach.

Keeping these apart is most of what analytical literacy consists of. It also keeps the reasoning honest in the other direction: a complete, coherent chain of custody is evidence about documentation quality, and does not by itself establish anything about safety or suitability for any purpose.

A traceability checklist

Terminology used above is defined in the research glossary, and the document-level reading skill is covered in how to read a COA.