How to read a certificate of analysis

A certificate of analysis is useful when you can connect it to the exact material and lot in front of you. A polished PDF on its own is not enough. The report needs to identify the sample, show who analyzed it, explain the method, and state what the reported numbers actually measure.

This guide gives researchers a repeatable way to review those fields without asking a COA to prove more than it can.

Scope note: This resource is about laboratory documentation for research materials. It does not provide dosing, administration, diagnosis, treatment, efficacy, or human- or animal-use guidance. Stack Research materials are for qualified in-vitro and laboratory research only, not for human or veterinary use.

Start with the record match

Before reading a result, make sure you have the right document.

Compare the report with the product or inventory record and check:

• Public product or sample name
• Exact configuration or strength
• SKU, when shown
• Lot, batch, or sample identifier
• Report ID
• Laboratory name
• Report date or completion date

The lot or batch field is one key connection. Confirm it against the sample description, configuration, laboratory sample ID, and report ID; no single field proves the match. A result associated with one lot does not establish the same result for another lot, even when the public product name and configuration look identical.

If the document has no lot, batch, or sample identifier, record that limitation. Do not fill the gap with a product-page claim or a different report.

Identify the laboratory and report

A usable record should make its source clear. Look for the laboratory’s legal or public name, contact information, a report number, dates, and an authorized signature or approval field.

Some laboratories also provide a verification code, QR code, or online report lookup. Use it when available. The information in the verification result should agree with the file you were given, including the sample name, report number, dates, and reported values.

A matching lookup helps show that the laboratory’s own system serves the same record. It does not verify the submitted sample, chain of custody, or any test outside the report. If the laboratory reported identity and purity, a valid lookup does not turn that report into sterility, contaminant, quantity, or suitability evidence.

If a report represents the laboratory as accredited, verify the facility and the relevant scope through the responsible accreditation body. The ILAC signatory search can help locate that body; a laboratory’s general accreditation does not show that every method is within scope.

Source check

1. Is the laboratory named?
2. Is there a report ID or verification code?
3. Can the laboratory be contacted through an independently found website or phone number?
4. Does the report identify an authorized reviewer or signature?
5. Does the verification record match the document exactly?

Read the sample description literally

The sample description tells you what the laboratory says it received. It may contain a public product name, internal code, physical description, or client-provided identifier.

Treat that field as a label, not a conclusion. A sample name does not establish identity by itself. Identity evidence comes from the stated analytical method and result.

Also check whether the report distinguishes among:

• The submitting company
• The sample name
• The lot or batch
• The tested configuration
• The laboratory’s internal sample number

These identifiers should form a consistent chain. If two fields conflict, stop and request clarification rather than choosing the one that seems more convenient.

Separate identity, purity, and measured content

These values answer different questions. They should not be used interchangeably.

FDA Q2(R2) treats identification, impurity or purity, and assay or content as distinct analytical attributes tied to the stated procedure.

Identity

Identity testing addresses whether the submitted sample met the report’s identity criteria under the stated method. Depending on the material and method, the report may compare an observed signal, mass, retention behavior, or spectrum with a reference.

An identity result does not automatically establish:

• Purity
• Total quantity or fill amount
• Sterility
• Absence of contaminants
• Suitability for a particular experiment
• The identity of another lot

Purity

A purity result reports the proportion calculated under the stated method. Check whether the report uses chromatographic area percent, mass fraction, or another basis; those quantities are not interchangeable. A chromatographic purity value is tied to that method’s conditions and detection limits.

Purity does not tell you how much material is present in a container. It also does not prove that every possible contaminant was tested or absent.

When reviewing purity, note:

• The analytical method
• The reported percentage
• Whether the result is area-based, mass-based, or otherwise defined
• Any reporting threshold or limit of quantitation
• Whether the report describes the calculation

Measured content

Measured content, assay, or quantity addresses how much of a stated analyte the laboratory measured in the submitted sample. Units matter. Milligrams, micrograms, percent by mass, and concentration are not interchangeable.

Compare the reported unit with the report’s sample preparation and calculation notes. If a value is reported with uncertainty, keep the uncertainty attached to the result rather than quoting only the central number. NIST’s traceability policy explains why traceability belongs to a measurement result and depends on a documented chain and stated uncertainty.

Additional panels

A report may include residual solvents, metals, bioburden, endotoxin, sterility, water content, or another specialized panel. Represent a panel only when the source document expressly includes it.

“Not detected” does not mean that an analyte cannot exist at any level. It means the method did not detect it under the stated conditions and reporting limits. Check the limit of detection or quantitation when provided.

Check the analytical method

The method tells you how the laboratory reached the result. The report may name an instrument or technique, reference a standard procedure, or provide an internal method code.

A method field should be specific enough to distinguish one type of evidence from another. Generic language such as “lab tested” is not a method.

Record:

• Technique or instrument
• Method or procedure identifier
• Reference standard, if stated
• Sample-preparation notes
• Reporting or quantitation limits
• Any stated deviations or qualifications

You do not need to become an analytical chemist to perform this review. The practical question is simpler: does the report state a method that is relevant to the result being quoted?

Follow the dates

Reports often show several dates:

• Date received
• Date prepared
• Date analyzed
• Date completed
• Date issued
• Date revised

These dates describe the report’s chain of events. They are not automatically an expiration schedule.

If a document was corrected or reissued, look for a revision date and an explanation. The current public record should make it clear which version is active. Replacing a file without identifying the change makes the history difficult to audit.

Keep reported results attached to their limits

Every result has a scope. That scope may be defined by the sample, lot, method, instrument conditions, reporting threshold, uncertainty, or a note in the report.

Preserve those limits when summarizing the document. Avoid statements such as:

• “The product is completely pure.”
• “The entire inventory is verified.”
• “The COA proves the material is safe.”
• “The report confirms sterility” when no sterility panel appears.
• “The laboratory approved the product” when it only reported analytical results.

A more accurate summary identifies the lot, laboratory, method, date, and exact result without adding a broader conclusion.

Review the document for internal consistency

Before accepting the record, compare the fields with one another.

Check for:

• A sample description that matches the public record
• A consistent lot or batch identifier
• Units that remain consistent between result and conclusion
• Dates in a plausible sequence
• A report ID that matches the verification lookup
• Page numbers and report identifiers on multi-page files
• A signature or approval field on the final version
• Notes or qualifiers that have not been cropped out

A screenshot of one result line is not a substitute for the full report. The omitted header, method, footnotes, and signature may contain the information needed to interpret it.

Common warning signs

Pause the review when you find:

• No laboratory name or source
• No report ID
• No lot, batch, or sample identifier
• A generic COA reused across multiple lots
• A result quoted without its unit
• A purity value presented as measured quantity
• An identity result presented as a contaminant or sterility panel
• A cropped image instead of the complete report
• Conflicting sample names, lots, dates, or report numbers
• A verification link that does not reproduce the same record
• Missing pages or footnotes
• Edits that obscure the original source

A warning sign does not always prove that a document is false. It means the record needs clarification before it can support a decision.

A five-point first-pass review

1. Match. Confirm the product or sample, exact configuration, and lot or batch.
2. Source. Identify the laboratory, report ID, dates, verification information, and authorization.
3. Method. Write down which method supports each result.
4. Scope. Separate identity, purity, measured content, and any additional panels. Record missing fields as unknown.
5. Preserve. Save the full source file, its public URL, the date accessed, and any verification result. If the record changes, retain the revision history.

How Stack Research publishes lot documentation

Stack Research keeps available records in its COA center and explains the publication and correction workflow on the Testing and Documentation page.

A Stack record marked Published means the approved document is available through Stack’s publication workflow for the record shown; it does not mean that a laboratory approved the material. Documentation pending means the required public evidence is not ready. A blank field stays unknown, and a report for one lot is not used as evidence for another.

Download the companion COA and Lot Verification Checklist to review a report field by field and keep a dated research record.

Frequently asked questions

Does a COA prove that every unit in inventory has the same result?

No. The result applies to the submitted sample. The report may associate that sample with a stated lot, but it does not establish uniformity across every unit or a result for another lot unless a documented sampling plan supports that conclusion.

Are identity and purity the same measurement?

No. Identity addresses what the sample is under the reported method. Purity addresses the measured composition under its reported method and calculation.

Does purity state how much material is in a container?

Not by itself. Measured content or quantity is a separate result with its own method and units.

Does “not detected” mean zero?

Not necessarily. It means the method did not detect the analyte under its stated conditions and reporting limits.

Can a report be verified online?

Some laboratories provide a lookup page, QR code, or verification code. Compare the retrieved record with the full report. A matching lookup helps show that the laboratory’s own system serves the same record. It does not verify the submitted sample, chain of custody, or any test outside the report.

What should happen when a report is corrected?

The replacement should remain attributable to its source and identify the revision or replacement date. Keep enough history to determine what changed and which version is current.

Does a laboratory report provide human-use guidance?

No. A laboratory report does not provide dosing, administration, diagnosis, treatment, efficacy, or human- or animal-use guidance.