Flower testing asks a relatively contained question: what is in this plant material. Cannabis edibles testing asks something harder — whether every unit in a batch contains what the label says, in a matrix designed to make chemistry difficult.
Producers moving from flower into infused products are frequently surprised by how differently the analysis behaves. Here is why.
Homogeneity Is the Central Problem
With flower, a potency figure describes the sample. With edibles, the figure has to describe every unit a customer might pick up.
If a batch of 1,000 gummies is labelled 10 mg per piece, the acceptable outcome is not an average of 10 mg. It is that any individual gummy is close to 10 mg. A batch averaging 10 mg while ranging from 3 mg to 22 mg is a labelling failure and a consumer safety problem, and averages conceal exactly that.
Homogeneity failures come from the production side: distillate that was not fully dispersed before depositing, emulsions that separated, temperature gradients in a mixing vessel, or cannabinoids settling during a long fill run. Testing detects the outcome. Only process control prevents it.
This is why homogeneity testing typically samples multiple units from across a production run rather than pulling a single composite. The variation between units is the measurement.
Matrix Effects
“Matrix” is everything in the sample that is not the compound being measured. In flower the matrix is plant material — well characterised and reasonably consistent. In edibles it is chocolate, gelatin, sugar, oils, emulsifiers, flavourings and colourants, and it changes completely from product to product.
Two problems follow.
Extraction recovery. Cannabinoids are lipophilic. In a high-fat matrix such as chocolate or a butter-based baked good, they bind to the fat and resist extraction into the analytical solvent. A method that recovers 98% from flower may recover far less from chocolate, and the reported potency comes back low — not because the product is under-dosed but because the method did not retrieve what was there.
Interference. Matrix components can suppress or enhance the instrument signal, or co-elute with the compounds of interest and distort peak integration. Sugars, oils and colourants are all capable of this.
Both are solvable with method development and validation for the specific matrix. Neither is solved by assuming a flower method transfers. This is the single most important question to ask a laboratory before sending infused products: has this method been validated on my matrix?
Milligrams, Not Percentages
Flower potency is expressed as a percentage by weight. Edibles are labelled in milligrams per serving and per package, because that is the unit a consumer can act on.
That conversion introduces its own requirements. The laboratory needs an accurate unit weight, and where a package contains multiple servings, the per-package figure has to hold as well as the per-serving one. Regulatory caps are typically written in milligrams per serving and per package, so an error in either direction is a compliance issue.
Because the numbers are small, precision matters more. A one percentage point error on 20% flower is proportionally modest. A 2 mg error on a 5 mg gummy is 40%.
Contaminants Concentrate Through the Chain
Edibles are made from extract, and extraction concentrates everything that partitions with the cannabinoids.
A pesticide residue sitting comfortably under the limit in source flower can exceed it in the distillate made from that flower, and therefore in the edible made from that distillate. The same applies to heavy metals. Testing the input material does not discharge the obligation to test the finished product, because the concentration relationship is not linear and not always predictable.
Solvent-extracted inputs also bring residual solvents into the finished product. And edibles introduce contamination routes flower does not have — food-grade ingredients with their own microbial and heavy-metal profiles, plus water activity in the finished product driving shelf stability. A full breakdown of the panels is in contaminant testing explained.
Stability Over Shelf Life
Cannabinoids degrade. THC converts to CBN with exposure to heat, light and oxygen, and the rate depends on the matrix and the packaging.
A beverage stored warm can lose meaningful potency over months. An emulsion can separate, which re-introduces the homogeneity problem after the product has already been released. A product that was accurately labelled on the day of testing can drift outside its stated range before it is consumed.
Stability testing — analysing the same batch at intervals across its intended shelf life — is how producers establish that their expiry dating is real rather than assumed. It is not universally required, and it is the difference between a defensible shelf life and a guess.
Beverages Are Their Own Category
Cannabis is not water-soluble, so beverages depend on emulsification to distribute cannabinoids through an aqueous matrix.
That makes them the most demanding infused product to test. Emulsions separate over time and with temperature change. Cannabinoids adsorb onto container surfaces, so a portion of the dose can end up on the inside of the can rather than in the liquid. Sampling technique matters more than in any other format — shaking or not shaking before sampling can change the result.
What Producers Should Do
- Validate the method for your matrix. Ask the laboratory directly. A generic method on an unusual matrix is where most surprises originate.
- Test homogeneity during development, not only at release. Sample units from the start, middle and end of a run to see what your process actually produces.
- Fix mixing before adjusting the recipe. Most homogeneity failures are process problems. Adding more distillate to correct a low average makes the spread worse.
- Test the finished product, not just the input. Concentration through the chain is why input results do not transfer.
- Run stability if you are claiming a shelf life. Otherwise the expiry date is unsupported.
- Account for the full package. Per-serving and per-package figures both have to comply.
We test the full range of infused formats, including gummies, chocolates, beverages, capsules and tinctures. The infused products page sets out which panels apply to each.
The short version: flower testing measures a plant, and edibles testing measures a manufacturing process. Treating the second like the first is where most problems begin.