A certificate of analysis is often treated as the end of a quality check. A PDF arrives, the reported value looks acceptable, and the document is filed with the purchase record. That may be enough to complete a procurement task, but it is not enough to preserve a defensible laboratory record, advises Samuel Desjardins.

The useful question is not simply, “does a COA exist?” but also, “can this report be tied to the material in front of us, and can that connection still be reconstructed later?”

That distinction matters because a COA is a source record, not a trust badge. Its value depends on continuity between the material, the sample, the test, the report, and the stored copy.

Before reviewing a purity value or identity statement, check whether the report identifies the material unambiguously. At minimum, the name, batch or lot number, sample identifier, report identifier, laboratory, and test date should form a consistent chain.

Small mismatches deserve attention. A product page may show one lot while the linked report names another. A supplier label may use an internal batch code that does not appear on the laboratory report. A report may describe a compound correctly but omit the sample identifier needed to connect it to the received container.

None of those gaps proves that the result is wrong. They do mean the record cannot support a strong lot-level conclusion without more evidence.

Record what was tested and how

A numerical result is easier to interpret when the report states the method, units, test date, and relevant detection or quantitation limits. The report should also make clear which attributes were tested.

For example, an assay result does not automatically establish identity. Identity evidence does not establish sterility. A high reported purity does not prove that the material matches a particular fill mass, storage condition, or unopened container. Each claim should stay inside the boundary of the method and sample described in the report.

This is a basic but often missed control. Procurement teams should separate the claims made in sales copy from the narrower evidence contained in the analytical record.

Preserve the original record

A screenshot is useful for quick reference but weak as an archive. It can omit page numbers, report identifiers, laboratory details, or revision information. It is also difficult to compare with a later version.

Keep the original file whenever possible. Record the source URL or delivery route, retrieval date, filename, and report identifier. A file hash can add another simple control by showing whether the archived copy has changed.

This does not make the report true. It preserves the exact record that was reviewed, which is a different and necessary function.

Add change control

Web-hosted reports can be replaced without warning. A link that still works may serve a different file next month. If the batch record depends on that report, silent replacement breaks continuity.

A practical change-control record should capture the previous file, the new file, the date the change was discovered, the stated reason, and the fields that changed. The reviewer can then decide whether the update is clerical or whether it affects an acceptance decision.

Version history matters most when the revised document changes the lot number, sample ID, method, laboratory, result, or test date. Those changes should trigger an impact review rather than an automatic overwrite.

Connect the COA to receiving controls

Document review works best when it joins the physical receiving process. The purchase record, shipment, container label, internal sample ID, storage location, and archived COA should point to one another.

A simple receiving workflow can record:

1. Supplier and purchase reference.

2. Material name and received quantity.

3. Supplier lot and internal sample ID.

4. Container condition and label details.

5. COA report ID, laboratory, method, and test date.

6. Storage requirements and assigned location.

7. Acceptance decision, reviewer, and date.

8. Exceptions, follow-up evidence, and final disposition.

This turns the COA from an isolated attachment into part of a traceable record.

Treat missing data as a documented state

Teams sometimes fill gaps with assumptions because the alternative feels incomplete. A better approach is to record “not provided,” “not verified,” or “requires clarification” at the field where the evidence stops.

That language is not a failure. It prevents a weak document from gaining authority as it moves through a workflow. It also gives the supplier or laboratory a precise question to answer.

Use a repeatable preflight

A short, consistent preflight is more reliable than an unstructured visual check. Before accepting a COA into the record, confirm:

- material, batch, and sample alignment;

- laboratory identity and report identifier;

- method, units, and test date;

- scope of the claims supported by the test;

- continuity between the received label and the report;

- original file preservation;

- version or replacement status;

- exceptions and reviewer decision.

The goal is not to turn every purchase into a forensic investigation. It is to make the important checks visible, repeatable, and reviewable.

A COA can be useful evidence. It becomes stronger when the laboratory can show what it represents, where it came from, which material it belongs to, and whether the record changed. Without that trail, the PDF may still look complete while the underlying decision remains difficult to defend.

Samuel Desjardins works on technical documentation, source record structure, traceability and publishing systems at LynxLabs Partnerships, Canada