How to Read a Certificate of Analysis Without Stopping at THC

A cannabis certificate of analysis can look like a product scorecard. It is better understood as a laboratory report tied to one representative sample from one production batch under one jurisdiction’s rules.

California’s Department of Cannabis Control requires every batch of cannabis goods to be tested before retail sale. Licensed laboratories report the results on a certificate of analysis, or COA, showing whether the batch passed or failed each required analysis. The same department requires laboratories to hold ISO/IEC 17025 accreditation, use standard operating procedures, maintain a quality-assurance program and participate in proficiency testing.

Those safeguards matter, but the first job belongs to the reader: establish that the report is the right report.

1. Match the identity block

The product name alone is not enough. A reader should compare the COA’s batch or lot identifier with the identifier on the package, along with the product type or matrix, producer information, laboratory name and relevant dates. California instructs distributors to confirm that the COA corresponds to the correct batch and is less than 12 months old before moving goods to retail.

Dates reveal the report’s sequence. A complete COA may show when the sample was collected, received and analyzed, and when the report was issued. A current-looking PDF tied to a different batch is not evidence for the package in hand. A QR code is only a delivery mechanism; the identifiers behind it still have to match.

The matrix also matters. Flower, infused pre-rolls, concentrates, beverages and edibles present different analytical challenges. California’s standardized cannabinoid method is mandatory for dried flower and non-infused pre-rolls, but not for infused pre-rolls and other manufactured matrices. A method name is meaningful only in relation to the material it was validated to test.

2. Read the overall result, then every panel

California’s required testing list includes cannabinoids and terpenes, residual solvents and processing chemicals, pesticides, heavy metals, microbial impurities, mycotoxins, moisture content or water activity, and foreign material. The applicable panels depend on the product and state rules.

“Pass” is therefore not a synonym for “nothing was detected.” It generally means the reported result met the jurisdiction’s specification for that analyte or panel. A contaminant may be present below an action limit and still receive a passing result. Another analyte may be reported as not detected because it was below the method’s detection capability. Those are different statements.

Readers should scan every required panel rather than treating the overall green check as the whole report. They should also look for missing pages, panels labeled not tested and footnotes that change which specification applies. A partial research or R&D panel is not automatically a regulatory-compliance COA.

3. Separate THC, THCA and total THC

The cannabinoid table may list acidic and neutral forms separately. Cannabis flower commonly contains much of its reported THC potential as tetrahydrocannabinolic acid, or THCA, alongside a smaller measured amount of delta-9 THC. “Total THC” is a calculation that combines them after accounting for the molecular-weight change associated with conversion of THCA.

Washington’s current laboratory rule states the formula directly: total delta-9 THC equals measured delta-9 THC plus 0.877 times measured delta-9 THCA. The 0.877 factor is not an extra potency bonus. It adjusts the acid-form mass to the neutral-form equivalent.

Units deserve the same attention. A flower result may be reported as percent by weight or milligrams per gram. Those units are convertible: one percent by weight equals 10 milligrams per gram. A manufactured product may instead report milligrams per serving, unit, milliliter or package. Comparing a percentage on flower with milligrams per package on an edible without converting the basis creates a false comparison.

The COA should also be compared with the retail label. California’s distributor checklist requires the labeled cannabinoid and terpenoid content to be consistent with the COA under the state’s labeling rules. A close batch match is more meaningful than a marketing number copied from a different lot.

4. Understand ND, LOD and LOQ

Laboratory abbreviations can make a zero look more certain than it is. California’s standard method defines the limit of detection, or LOD, as the lowest amount that can be distinguished from the absence of the analyte within a stated confidence limit. The limit of quantitation, or LOQ, is the lowest concentration the method can reliably quantify while meeting defined goals for bias and imprecision.

That creates three broad result zones. Above the LOQ, the laboratory can report a numerical concentration within the method’s range. Between the LOD and LOQ, the instrument may detect evidence of an analyte but cannot quantify it with the required reliability. Below the LOD, the report may say ND, not detected, or use another laboratory-specific notation.

ND therefore does not prove absolute absence. It means the analyte was not detected at or above the applicable reporting capability. A careful COA lists the method’s LOD and LOQ or provides them in a linked data package, ideally in the same units as the result.

5. Look for method and quality signals

The laboratory section should identify the analytical method and, on a detailed report, the instrumentation. California’s standardized flower method uses high-performance liquid chromatography and requires calibration standards, blanks, control samples, replicate analysis and matrix-spike work. It also requires laboratories to determine LOD and LOQ and to reanalyze samples associated with failed quality-control criteria.

These details are not decoration. A method blank checks whether sample preparation introduced contamination. Calibration verification checks that instrument response remains valid. A laboratory replicate evaluates precision. A matrix spike tests how the particular sample material affects measurement.

NIST’s Cannabis Laboratory Quality Assurance Program was created because comparability across laboratories cannot be assumed. The program sent common materials to participating laboratories, compared their measurements with accepted values and examined how preparation and analytical methods can introduce bias. An accreditation mark shows that a laboratory operates within a quality system; it does not mean all methods, matrices and measurements are interchangeable.

6. Remember what was actually sampled

A COA characterizes the submitted representative sample and, through the regulated sampling plan, supports a decision about the batch. It does not mean every package was individually tested.

California requires a laboratory employee to collect a representative sample from a batch held by a distributor. The distributor must observe the collection, ensure increments come from throughout the batch, preserve video and sign the chain-of-custody form. Once a sample has been taken, a second laboratory generally cannot resample that batch without department approval.

That system makes sampling part of the result. If the batch was not homogeneous, if identifiers were broken or if the sample was not representative, a precise instrument number can still describe the wrong thing. The best COA reading therefore follows the chain backward: package to batch, batch to sample, sample to method, method to result.

A one-minute trade check

For buyers, retailers and operators, a fast review can follow a fixed order:

• Match the batch or lot number and product matrix.
• Confirm the licensed laboratory and report dates.
• Read the overall status and every required panel.
• Compare THC, THCA and calculated total THC in the reported units.
• Check LOD, LOQ, ND and any action-limit footnotes.
• Confirm the package label is consistent with the batch report.
• Treat the COA as evidence about a regulated sample, not a complete verdict on manufacturing quality.

A COA is most useful when it is read as a chain of evidence rather than a potency menu. The headline number may attract attention, but identity, sampling, method scope and contaminant panels determine what the document can actually support.