Batch, Lot, Sample ID, and Test Date: Matching a COA to a Product

A certificate of analysis can contain numbers and answer the wrong question. Before reading results, confirm that the report belongs to the product and batch in front of you.

Build an identity chain:

product and variant → package batch or lot → laboratory sample → report → tested matrix and dates

If one link is missing, the report may still be informative about an ingredient or earlier production stage, but it should not be presented as a matched finished-product result.

The identifiers do different jobs

Field	Usually identifies	Main question
Product name	Commercial item or formulation	Is this the same named product?
Variant	Strength, size, flavor, count, or format	Is this the same version I selected?
Batch or lot	Defined production quantity	Does the package code match the report?
Sample ID	Material received or created in the lab system	Can the lab trace the tested portion?
Regulatory UID	Record in a government track-and-trace system, where used	Does the regulated record connect to the sampled batch?
Matrix	Physical material tested	Was it flower, oil, gummy, capsule, isolate, or another material?
Date collected	When the sample was taken	When did sampling occur?
Date received	When the lab accepted it	Was there a recorded transfer into laboratory custody?
Analysis date	When a test was performed	When were the measurements made?
Report or release date	When the document was issued	Is this the original or a later report?

Laboratories and regulatory programs do not use every term in exactly the same way. Some records use “batch,” others “lot,” and some display both a client lot number and an internal laboratory sample ID. Do not assume that two different codes are errors until the document explains what each code identifies.

Begin with the package, not the PDF

Locate the code on the actual package. It may appear on the label, bottom, cap, heat seal, pouch seam, or an applied sticker. Record it exactly, including letters, leading zeros, dashes, and suffixes.

Also record:

• full product name;
• selected strength or concentration;
• package size or count;
• flavor or formulation, when relevant;
• product format;
• any manufacture, package, best-by, or expiration date; and
• QR or direct document link.

A difficult code should be photographed rather than reconstructed from memory. O and 0, I and 1, or a missing suffix can point to different records.

If the seller supplies a general “latest COA” without a package identifier, ask which lots it covers and how that connection is documented.

Find the client identifier and the laboratory identifier

On the report, look for labels such as:

• batch number;
• lot number;
• client sample ID;
• laboratory sample ID;
• sample name;
• source batch;
• METRC or other UID;
• work order; or
• accession number.

The package code should match the report’s product-side batch or lot field—not merely the lab’s internal sample number. A laboratory sample ID can be unique and valid while identifying a sample from a different production lot.

California’s regulated cannabis rules provide a clear example of the distinction. The chain-of-custody record includes the batch number and an assigned unique sample identifier. The COA then identifies the batch, matrix, unique sample identifiers, and sample history. For products already packaged at sampling, the rules require the labeled package batch number to match the COA batch number.

That state system is not a universal template for retail hemp reports. It shows why both the production identifier and the laboratory identifier matter.

Confirm the matrix

The matrix is the kind of material tested. It may be:

• dried flower;
• biomass;
• crude extract;
• distillate;
• isolate;
• carrier oil;
• finished tincture;
• gummy;
• capsule;
• topical; or
• another finished or intermediate material.

Matrix determines what the result can describe. A flower report does not establish the composition of oil produced later. A distillate COA does not test the carrier oil, flavoring, capsule shell, gummy base, or final unit-to-unit distribution. An ingredient result may support supplier qualification, but it is not automatically a finished-product result.

The report’s units should fit the matrix. Percent by mass or milligrams per gram may be natural for a concentrate. A finished oil label may require conversion using bottle volume, density, or total mass. A gummy claim stated per piece cannot be verified from a per-gram result without the tested unit mass and a sound sampling basis.

Potency Units Explained covers those conversions after identity is established.

Read the dates as a sequence

One “test date” can hide several events. A robust report may show:

1. sample collection date;
2. laboratory receipt date;
3. preparation date;
4. analysis date for each method;
5. review or approval date; and
6. report issue or revision date.

These fields are not interchangeable. A report issued in July may contain measurements made in June from a sample collected in May.

Compare the sequence with the product record. A sample collected before the stated production event, or a report issued long before the lot existed, needs an explanation. A report date later than purchase is not necessarily improper—it could be a corrected report—but the revision history should identify what changed and preserve the connection to the original sample.

Do not reject a report solely because it is old; a seller may still be distributing that exact lot. The decisive questions are whether the package matches the lot and whether the report describes the relevant material. Age matters when the formula, supplier, package, or batch has changed.

Separate a crop-lot result from a finished-product COA

USDA’s Domestic Hemp Production Program uses “lot” for crop acreage designated by a producer. Its compliance testing is designed to measure total delta-9 THC in a composite hemp sample on a dry-weight basis. USDA reporting includes producer and license information, the FSA-issued lot identification number, test date, pre- or post-harvest status, and result.

That is valuable agricultural compliance evidence. It does not automatically test a finished retail oil, gummy, capsule, or isolate made after harvest.

Ask what stage the report covers:

• pre-harvest crop;
• harvested flower or biomass;
• extracted ingredient;
• refined ingredient;
• bulk formulation; or
• packaged finished product.

Several reports may document one product’s history. They answer different questions and should not be collapsed into a single “lab tested” claim.

Check who created and approved the report

Identify:

• laboratory name and address;
• laboratory license or accreditation information, when applicable;
• client or submitting party;
• method or method reference;
• signature or electronic approval;
• report status;
• page count; and
• document or revision number.

Then verify the laboratory through the relevant regulator or accreditation body when the jurisdiction provides a directory. A logo on a PDF is not verification.

An authentic report can still be limited. Accreditation may cover some methods and matrices but not others. Hemp Lab Panels Explained shows how to identify the panels actually reported.

Handle amended and reissued reports carefully

Corrections happen. A laboratory may fix a client name, sample description, transcription error, or other field. The revised report should make the history visible through a revision number, amended status, new issue date, reason, or reference to the superseded document.

Compare the entire report, not just the corrected field:

• Does the sample ID remain the same?
• Did any result change?
• Did the method or reporting limit change?
• Is the signature current?
• Does the report state that it supersedes an earlier version?
• Does the QR code now resolve to the revised report?

California’s 2026 rules, for example, require supervisory validation and a signed and dated COA, and they treat changes to reported results differently from minor corrections. Other jurisdictions and private reports may use different procedures.

QR codes are routes, not evidence

A QR code can make retrieval easier. It does not establish that the destination is complete, permanent, or matched.

After scanning:

1. confirm the domain belongs to the laboratory, regulator, or identified document host;
2. confirm the batch or lot against the package;
3. compare the product and matrix;
4. check every page is present;
5. note whether the link points to a document, viewer, or changing “latest report”; and
6. save the direct report link and issue date when a durable record matters.

A cropped image of a potency table is not a complete COA. It can omit identity fields, methods, reporting limits, signatures, footnotes, or failed and untested panels.

Common mismatch patterns

Pause and request clarification when:

• the package has a lot code but the report does not;
• the report names a different strength, flavor, or size;
• the matrix is an ingredient rather than the finished product;
• only a lab sample ID is shown and no production code connects it to the package;
• the dates occur in an impossible or unexplained order;
• a report combines several products without explaining the sampling design;
• a QR code opens a different product or only the newest batch;
• the product label changed but the same undated report remains attached;
• pages, signatures, or method notes are missing; or
• the file appears edited outside the laboratory’s document system.

Do not repair the match by assuming that visually similar products share a batch.

The matching checklist

Before reading the result table, answer yes, no, or unknown:

• Does the package product name match?
• Does the selected variant match?
• Does the package batch or lot match the report’s production identifier?
• Is the laboratory sample ID visible?
• Is the tested matrix the material the claim describes?
• Are collection, receipt, analysis, and report dates clear?
• Is the laboratory identifiable and verifiable?
• Is the report complete and approved?
• Is any amendment or revision explained?
• Does the unit basis fit the format?

Unknown is not automatically failure. It is a reason to obtain the missing record before making a batch-specific claim.

How to Read a Hemp Certificate of Analysis covers the result table, limits, methods, and sample scope. Identity comes first: the best analysis cannot describe a product it was never connected to.

Primary sources

• USDA AMS, Laboratory Testing Guidelines
• USDA AMS, Hemp Frequently Asked Questions
• California Department of Cannabis Control, Medicinal and Adult Use Cannabis Regulations, revised January 1, 2026
• FDA/ICH, Q7 Good Manufacturing Practice Guidance for Active Pharmaceutical Ingredients