Cannabis COAs are a silent quality crisis

Most cannabis product developers treat a certificate of analysis (COA) as the gold standard. This is how they verify purity, potency and, in a perfect world, product safety. But behind the clean-looking lab report, something more complicated is unfolding.

Across the supply chain, a quiet quality crisis is brewing. Cannabinoid certificates of analysis (COA) have been selectively framed and outright manipulated in the past. Ingredients that look compliant on paper are inconsistent in practice, and product developers are paying the price in reformulations, delays, and broken consumer trust.

Advertisement

COAs reveal only what labs are asked to test for

The deeper issues? The COAs usually only reveal what labs are asked to test for. Even the most clean COAs can leave out factors that could impact consistency, potency and quality. Some cannabis ingredient producers are well-versed in how to pass COAs without offering a pure product. A supplier can remain technically compliant if they test only a part of the lot, use narrow detection thresholds or leave entire classes of impurities untouched. It is exacerbated by the fact labs who consistently report higher potencies or fewer failures for residual solvents, or other contaminants, will get repeat business. Even reputable brands are susceptible to being misled. This lack of visibility can have a negative impact on brands. For brands, it’s a threat to consistency and consumer trust.

Consistency is the real promise brands must keep

Consider Coca-Cola: The formula is designed to deliver the same taste and experience in every can someone cracks open anywhere in the world. Cannabinoid products are expected to meet the same standards. Even if test results are clean, consistency can be compromised when ingredients differ from batch to batch, especially in products that are based on sensory or effect-driven attributes, such as gummies, tinctures and topicals. The COA of the supplier promised high purity and dosage verification. Within weeks, however, the customer complaints began to pour in. Others reported outcomes that did not match the product’s promises. After third-party analysis, the brand asks their lab to test for both delta-9 tetrahydrocannabivarin (THCV; the cannabinoid the plant makes) and delta-8 THCV (an isomer that is the signature of degradation and/or suboptimal production techniques). The results show that the batch is mostly made up of delta-8 THCV. Consequently, the brand is forced to pull product off shelves, source new inputs, rerun tests, and re-market the finished product, which costs the team weeks of time, tens of thousands of dollars, and immeasurable reputation damage.

And it’s not just consumer-facing brands that feel the impact.

Real-world variability has real-world costs

One example distributors could face is inconsistency in batch production. Distributors could face inconsistency in batch production after a third-party has verified a successful test batch. The product initially works well, and it is sent out to several formulator partners. The second shipment, which is labeled the same as the first batch, shows clear differences in color and solubility. Partners are forced to pause launches, run tests again, and delay production. The COA was identical for all shipments but the material was clearly not.

A structural issue is part of the problem. In terms of mechanism-first research, the industry is still underfunded. Instead of asking how cannabinoids work — and how different inputs affect biological pathways — most companies focus on outcome-first claims that validate what’s already being sold.

A research gap is limiting progress

“Mechanism-first research gives us a blueprint,” said Dr. Tony Ferrari, an analytical chemist with deep expertise in cannabinoid formulation. The research that focuses on the outcome is a sales pitch.

This leads to a scientific blindspot, which affects everything from testing standards to consistency of ingredients. Without fully understanding how minor cannabinoids interact with the body or degrade during processing, we miss clues that could inform better safety protocols and formulation techniques.

In most industries, research starts in the lab, but “in cannabis, the reverse happened,” said economist Dr. Ruth Fisher. Researchers began to study the effects of cannabis in the real-world after people started using it in large quantities. We can discover patterns and side-effects that we would not have noticed in a laboratory setting. It will require more than just more accurate COAs. It will require:

Broader impurity panels and updated reference standards.

Transparency from labs about what

isn’t

being tested.

Redundant COAs from at least two different labs for the same batch of material (to demonstrate results are consistent).

Ingredient partners that guarantee consistency batch to batch and ship the actual product represented on the COA.

Open collaboration across brands, labs, and researchers to share real-world observations.

COAs should never be the finish line. In a space as complex and rapidly evolving as cannabinoids, a COA should be the starting point, providing a baseline that invites deeper questions, better research, and long-term trust-building.

• Because what’s not listed on your COA could end up being what costs your brand the most.
• FAQ: COA Challenges and SolutionsWhy isn’t a COA enough to confirm cannabis ingredient purity?Because COAs only show what the lab was instructed to test for — and reporting limits may exclude relevant impurities or isomers.
• How can a brand verify ingredient consistency across batches?
• Request redundant COAs from two labs and require suppliers to guarantee batch-to-batch equivalency.
• What impurities are most commonly missing from standard COA panels?

Isomers, synthesis byproducts, residual intermediate compounds, terpene-derived artifacts, and degradation markers.

How can labs reduce the incentive bias that leads to inflated potency reports?

Transparent methodologies, standardized reference materials, and third-party auditing reduce incentive pressure.