Laboratory Variability, Batch Numbers and Test Dates

Why results differ

A cannabis lab result is a measurement made on a small sample by a specific method on a specific day. Several things can move the number: how the sample is taken and ground, how the compounds are extracted, whether the sample is heated, which instrument and settings are used, and how the lab calculates totals. A review of cannabinoid analysis describes the task as tricky, noting that the instability of cannabinoids, variability in sample preparation, extraction and analysis, and isomeric forms complicate the picture, and that each published method has inherent limitations. [1]

What studies have found

• Interference can inflate THC. NIST researchers reported chromatographic interferences, including from CBNA and synthetic delta-8 THC by-products, that can inflate delta-9 THC levels and lead to misidentifying hemp as marijuana. They also showed that adjusting the method can resolve the problem. [2]
• Labels versus retest. In a study of 23 retail flower samples from 10 Colorado dispensaries, average observed THC was 14.98 +/- 2.23%, 23.1% lower than the lowest label values. About 70% of samples were more than 15% below the label. [3] It is a small sample from one region, so it shows a pattern of concern, not a universal rate.
• Larger check. In 277 Colorado flower and concentrate products, observed THC was significantly lower than labeled for both. Within 15% of the label were 96.0% of concentrates and 56.7% of flower. [4] One author group was employed by a licensed cannabis company, which the paper discloses.

"Lab shopping" and potency inflation

The Colorado retail study says reports of inflated THC potency and "lab shopping" to obtain higher potency results had been circulating, and that THC potency is a primary factor in pricing flower. [3] The authors cite a lack of standardized testing protocols, limited regulatory oversight and financial incentives to market high THC as likely contributors. [3] The 2025 study calls for independent and blinded verification of labels and testing procedures. [4] This site does not name labs or products and does not rank them.

Batch numbers

A batch (or lot) number identifies a particular production run. A lab report belongs to that run, not to a product name. Health Canada's label guide says a product label will contain the lot number and the date the product was packaged. [5] The USDA hemp rule requires that samples from one lot not be commingled with hemp plant material from other lots, which shows why the lot is the unit that testing describes. [6] If the batch on the report does not match the batch on the package, the report does not describe your product.

Test dates

• A result describes the sample at the time of testing. Cannabinoid amounts can change as material ages, and one analytical review lists instability of cannabinoids among the complications in measurement. [1]
• Look for the date tested and the date reported, and compare to the packaging date if the label gives one. [5]
• An old report for a new batch is a warning sign, not proof of anything.

A fair checklist

1. Does the batch number on the report match the package?
2. Are the test and report dates present, and recent relative to packaging?
3. Is the method and lab named, with limits and uncertainty? See limits of detection and quantification.
4. Are totals explained? See total THC and total CBD.
5. Does the report prove anything about effects? No; see why a COA does not prove effectiveness.

Start with how to read a cannabis COA for the whole report.