How to Read a Certificate of Analysis (COA)

A certificate of analysis looks difficult because it places several kinds of information on one page: identity, chemistry, units, methods, dates, and pass-or-fail language. The fastest way through it is not to start with the biggest number. Start with the identity block. A precise result for the wrong product is still the wrong document.

This guide reads a COA in the order that makes it useful: match the sample, identify the laboratory, locate the panel, understand the units, and compare the result with the claim being made. It also separates what a report can establish from what it cannot.

General information, not a substitute for the advice of an attorney. Jurisdiction: Texas, with United States federal law where noted. Information current as of August 2026.

What a certificate of analysis is

A certificate of analysis, or COA, is a report issued by a laboratory for an identified sample. It records the panels performed and the results produced under the laboratory’s methods. In Texas, the consumable-hemp rules require a batch COA from an ISO/IEC 17025 accredited laboratory, and the COA has to report the percentage of delta-9 THC, total delta-9 THC, and total THC per container (25 TAC §300.301(d)(12); Executive Order GA-56).

That legal requirement does not make every COA interchangeable. The document still belongs to one sample. Its value depends on whether the sample identity, product, batch, date, and panel match the question being asked.

Step one: match the product before reading the chemistry

The identity block usually sits near the top. It can include the client or brand, product name, sample name, matrix or product type, batch or lot identifier, and dates associated with receipt, testing, and reporting.

Four matches matter:

Brand or client. The company named on the report corresponds to the product being examined.
Product name and form. A tincture report is not a gummy report; one flavour, strength, or package format does not silently stand for another.
Batch or sample identity. The identifier connects the tested sample with the relevant production run.
Date. The report belongs to the current formulation and a relevant batch rather than to an earlier version with the same name.

A mismatch at this stage ends the analysis. Chemistry cannot repair identity. A report can be authentic and analytically sound while still being irrelevant to the package in hand.

Step two: identify the laboratory

The report names the laboratory and commonly includes contact information, a report or sample number, signatures or approval marks, and accreditation information. Texas requires consumable-hemp batch testing through an ISO/IEC 17025 accredited laboratory (25 TAC §300.301; Texas State Law Library cannabis guide).

Accreditation and independence answer different questions. Accreditation concerns the laboratory’s competence for the scope identified by its accreditation. Third-party status means the laboratory is separate from the company selling the product. Neither phrase turns one report into proof of every possible quality claim.

A laboratory logo on its own is not the useful part. The useful elements are the identified lab, the traceable report number, the sample information, the methods, and the results. A screenshot cropped so tightly that those elements disappear removes much of the document’s value.

Step three: find the panel that answers the question

A COA may include one panel or several. Potency, heavy metals, pesticides, residual solvents, microbiological screening, and other analyses are different tests. A result appears only for the panel the laboratory actually performed.

The panel heading matters because the words “tested” and “passed” have no universal scope. “Passed potency” does not mean “passed every contaminant panel.” “No result reported” does not mean zero. A blank panel does not become a negative result by implication.

For cannabinoid potency, the analyte table lists named cannabinoids with results in one or more units. The row names determine what was measured. The total row is a calculation or reported aggregate, not a substitute for reading the components when the legal question turns on a particular cannabinoid.

Step four: read the units before reading the number

The same sample can be expressed as a percentage, milligrams per gram, milligrams per unit, or milligrams per container. Those figures are not competing results; they answer different scale questions.

Percentage by weight describes a proportion of the sample.
Milligrams per gram describes concentration by mass.
Milligrams per unit or serving applies the measured concentration to a defined piece or serving.
Milligrams per container applies the result to the package unit identified by the report.

Texas currently regulates consumable-hemp potency through an “acceptable hemp THC level” of total delta-9 at or below 0.3% dry weight, measured as (0.877 × THCA) + delta-9 (25 TAC §300.101(1), §300.302(b); Texas State Law Library cannabis guide). Texas also requires total THC per container to appear on the COA as a disclosure (25 TAC §300.301(d)(12)). The disclosure is not a Texas milligram ceiling; 25 TAC ch. 300 contains no per-serving or per-container milligram cap.

Federal law adds a different layer on 2026-11-12. Section 781 of P.L. 119-37 moves the federal hemp definition to total THC at 0.3% dry weight and caps finished products at 0.4 mg total THC per container (Congressional Research Service IF13136). as of August 2026 the FDA had not published a definition of “container,” so this article does not resolve multi-layer packaging or assert later agency activity (CRS IF13136).

Step five: understand ND, LOQ, and LOD

Laboratory tables often use abbreviations instead of a zero.

ND commonly means not detected under the method and reporting limits used.
LOD is the method’s limit of detection: the point at which the analyte can be distinguished from background.
LOQ is the limit of quantitation: the point at which the amount can be reported with the method’s defined quantitative confidence.

ND does not mean the laboratory proved absolute absence. It means the analyte was not detected at or above the applicable limit. A result below the LOQ can be detectable without being reliably quantifiable. The report’s notes or legend control the exact meaning used by that laboratory.

Step six: separate measured values from label claims

The package states what the product is represented to contain. The COA states what the laboratory reported for the tested sample. A useful review places those two figures beside each other without silently changing units or package definitions.

Three errors are common:

Comparing a per-serving label claim with a per-container laboratory result.
Comparing a percentage with milligrams without using the product mass.
Comparing a total-cannabinoid marketing number with one analyte row.

A match requires the same substance, the same unit, and the same package or serving basis. If any of those changes, the arithmetic has changed before the numbers are compared.

Step seven: check dates and batch relevance

A COA does not become false because it is old. It becomes evidence about an old sample. That distinction matters whenever a formula, supplier, package size, strength, or manufacturing batch changes.

Texas requires the product label to carry a batch number and batch date, along with a conspicuous URL reaching the COA in no more than three steps (25 TAC §300.402(a); Executive Order GA-56). Those identifiers create the route from package to document. Where the package and report cannot be connected, the report may still describe the product family while failing to establish the current batch.

FireBar Labs publishes consumable-product results by product name at /pages/lab-testing. Consumable packaging carries a QR code that opens that product’s lab results. Products do not carry a customer-facing batch-code search box, no printed COA ships in the parcel, and reports are not emailed with an order.

What a COA can establish

The identity information supplied for the tested sample.
The laboratory and report associated with that sample.
The panels performed and methods or method references shown.
The measured or reported values for the named analytes.
The reporting limits, qualifiers, and pass-or-fail status defined on that report.

What a COA cannot establish

Anything about a product or batch that does not match the sample.
A result for a panel that was not run.
How every unit in a batch changed after the sample was collected.
A universal claim that the product is “safe” in every use or for every person.
The legal classification of a cannabinoid where the law turns on identity or origin beyond the number reported.
How “container” applies under the federal 0.4 mg limit after 2026-11-12, because as of August 2026 the FDA had not published a definition of that term (CRS IF13136).

Authenticity and relevance are separate checks

A report can be authentic without being relevant. Laboratory contact details, report numbers, approval marks, and traceable identifiers help establish that the document came from the named laboratory. Product, sample, batch, form, and date establish whether that authentic document answers the question in front of the reader.

This distinction catches two opposite mistakes. Rejecting an old report as “fake” confuses age with authenticity. Accepting a genuine report for another product confuses authenticity with relevance. The accurate conclusion can be narrower: the document appears genuine, but it does not establish this package.

A cropped image makes both checks harder. Removing the header, footer, page numbers, notes, or signature block may leave the analyte table readable while stripping away the information that connects it to a lab and sample. The complete report carries context the largest number cannot.

The Texas three-step access rule

Texas labels must provide a conspicuous URL that reaches the certificate of analysis in three steps or fewer (25 TAC §300.402(a); Executive Order GA-56). A QR code can make that route fast, but the legal text is concerned with access to the document, not with the visual presence of a square pattern.

The practical route has three separate checks: the code or URL resolves; it lands on the relevant product; and the resulting document contains enough identity information to connect it back to the package. Our companion article explains what good COA access looks like on a package.

A compact reading order

Match brand, product, form, batch or sample, and date.
Identify the laboratory, report number, and accreditation information.
Locate the panel relevant to the claim.
Read the units and package basis.
Read qualifiers such as ND, LOD, and LOQ.
Compare like with like: same analyte, same unit, same serving or container.
Note what was not tested and what the document cannot answer.

That order prevents the largest number on the page from deciding the interpretation before the identity and units are known.

Frequently asked questions

What is a COA?

It is a laboratory report for an identified sample, listing the panels performed and the results produced. Texas requires a batch COA from an ISO/IEC 17025 accredited laboratory for consumable hemp (25 TAC §300.301; Texas State Law Library cannabis guide).

Does ND mean zero?

No. ND commonly means the analyte was not detected at or above the method’s applicable limit. The report legend defines the laboratory’s exact use of the abbreviation.

Does a passing potency test mean contaminant panels passed?

No. Each panel has its own scope. A potency result establishes only the potency findings reported for the sample; contaminant claims require the corresponding panels on the document.

Does Texas cap THC at a certain number of milligrams per package?

No. Texas uses a total-delta-9 percentage standard at or below 0.3% dry weight under 25 TAC §300.101(1) and requires total THC per container to be disclosed on the COA under §300.301(d)(12). The state rule contains no per-container milligram ceiling.

What changes federally on 2026-11-12?

Section 781 of P.L. 119-37 changes federal hemp to a total-THC test at 0.3% dry weight, caps finished products at 0.4 mg total THC per container, and excludes cannabinoids synthesized outside the plant (CRS IF13136). as of August 2026 the FDA had not published a definition of “container”; this article does not assert later agency activity.

Where are FireBar Labs COAs?

Results for consumable products are organized by product name on the lab-testing page. The QR code on consumable packaging opens that product’s results. Bag QR codes lead to setup instructions instead; bags are field-tested, not lab tested.

About this article

This article is general information about the law, published for educational purposes. It is not a substitute for the advice of an attorney. Laws change and vary by state — consult an attorney licensed in your jurisdiction about your specific situation. Jurisdiction covered: Texas, with United States federal law where noted. Information current as of August 2026.

Sources: the Texas consumable-hemp rules at 25 TAC ch. 300; Executive Order GA-56; the Texas State Law Library cannabis guide updated 2026-08-05; §781 of P.L. 119-37; and the Congressional Research Service overview of the federal hemp redefinition, IF13136. Published by The FireBar Labs Team.

Explore FireBar Labs

Questions about finding or reading a FireBar Labs result? Email support@firebarlabs.com.

How to Read a Certificate of Analysis (Without a Chemistry Degree)