---

Every product page on our site links to a COA. We mention COAs in our blog posts, our staff talks about them with customers at the counter, and we’ve built our entire sourcing process around them. So we figured it was worth writing down exactly how we read one — not the version where every section gets a paragraph and you come away understanding the vocabulary but not the judgment, but the version where we tell you what we’re actually looking for and what makes us put something down and move on.

This is how we evaluate a COA before anything comes into Sherlocks.

What a COA actually is — and what it isn’t

A Certificate of Analysis is a document produced by an independent, third-party laboratory that tests a specific batch of cannabis or hemp product and reports what it found. The key word is independent. The lab that tests the product should have no financial relationship with the brand selling it — they’re not there to produce a favorable result, they’re there to produce an accurate one.

What a COA isn’t is a guarantee. It’s a snapshot of one tested sample from one batch at one point in time. A COA from six months ago on a product that’s been sitting in a warehouse tells you something, but it doesn’t tell you everything. We look at COA dates as part of our evaluation. Fresh testing on fresh product is what we want to see from our suppliers, and it’s what we provide to our customers.

The lab itself matters too. We look for ISO/IEC 17025 accreditation on every COA we review. That accreditation means the lab has been independently audited for technical competency and that their methods and results meet an internationally recognized standard. If a COA doesn’t show accreditation information, that’s a reason to ask questions before it’s a reason to accept the product.

The cannabinoid panel — and why we don’t lead with THCa percentage

The cannabinoid panel is the section most people look at first. It shows the concentration of each detected cannabinoid as a percentage of the product’s dry weight — THCa, delta-9 THC, CBD, CBG, CBC, CBN, and others depending on the lab’s panel.

Most buyers go straight to the THCa number. We understand why, but that number alone tells us less than people think. A flower that tests at 32% THCa but was grown with synthetic nutrients, harvested early, and improperly cured will smoke worse than a living soil cultivar that tests at 24% and was given the time it needed. The THCa percentage reflects a lot of variables, and potency is only one of them.

What we look at in the cannabinoid panel is the full picture. A complete, well-grown plant shows minor cannabinoids alongside the THCa — CBG, CBC, and sometimes CBDa present in measurable amounts. That breadth reflects a plant that was allowed to develop fully. It’s not a rule without exceptions, but consistent presence of minor cannabinoids is a signal we weight positively. A cannabinoid panel that shows nothing but a THCa number and zeroes across the board is worth questioning.

The delta-9 THC number is the compliance number. Federal hemp law requires delta-9 THC to be below 0.3% on a dry weight basis. Every product we carry passes that threshold. As regulations evolve toward total THC calculations — which account for THCa converting to THC when heated, using the formula Total THC = delta-9 THC + (THCa × 0.877) — that’s a standard we’re already watching closely and building our sourcing around.

The terpene panel — the section most people skip

This is the one. If there’s a single section of a COA that separates a retailer who understands what they’re selling from one who doesn’t, it’s whether they even look at the terpene panel — and what they do with the information when they do.

Terpenes are the aromatic compounds that determine how a strain smells, how it tastes, and in significant part how the effect feels. They’re produced in the same trichomes as THCa, which is why living soil cultivation matters so much to terpene expression — the plant’s access to a full range of nutrients and trace minerals directly influences the depth and complexity of its terpene profile.

Total terpene content above 2% is a reasonable baseline for quality flower. Above 3% starts to indicate something genuinely exceptional. We’ve seen living soil cultivars from Seeds of Kismet test above 3% consistently, and you can smell the difference before you even finish opening the jar.

Beyond the total number, we look at the specific profile. Myrcene heavy with strong caryophyllene suggests a heavy, relaxing indica-leaning experience. Limonene and terpinolene forward usually means something more energetic and cerebral. A terpene panel dominated by a single compound with almost nothing else present is less interesting to us than one where five or six terpenes are competing for attention — that complexity is what produces the layered, strain-specific effects that experienced consumers are actually looking for.

If a COA doesn’t include a terpene panel at all, that tells us something. Not every lab runs them by default, and not every supplier pays for them. We ask for them. If a supplier can’t provide one, we want to understand why before we go further.

The safety panels — what you want to see is nothing

Pesticides, heavy metals, residual solvents, and microbial contamination are the four safety panels that matter most. What you’re looking for across all of them is ND — not detected — or results that fall well below the action limits set by your state’s regulatory framework.

Pesticide panels are especially important in the hemp market because hemp operates under less regulatory scrutiny than licensed cannabis in most states. Some operators take advantage of that gap. We don’t accept product with pesticide detections, period — not at any level.

Heavy metals testing covers lead, arsenic, mercury, and cadmium. Cannabis is a bioaccumulator, meaning it pulls metals from the soil it’s grown in. Living soil cultivation done correctly with clean inputs shouldn’t produce heavy metal detections, but we verify rather than assume.

Residual solvents matter most for concentrates and extracts rather than flower. For the live hash rosin we carry, this panel should be clean by the nature of the product — no solvents are used in solventless extraction, so nothing should be there to detect.

Microbial panels cover yeast, mold, E. coli, and salmonella. Flower that was improperly dried or stored can show mold counts that create real health risks, particularly for anyone with compromised immunity. This is one of the panels where we’re least flexible.

Batch number and date — the two things most consumers never check

Every COA should include a batch or lot number that corresponds to a specific harvest or production run, and a test date that tells you when that batch was analyzed. The batch number is how you connect the document to the actual product — if you buy THCa flower and the batch number on the jar doesn’t match the batch number on the linked COA, that’s a problem worth flagging.

The date matters because COAs have a practical shelf life. A test result from eighteen months ago on a product that’s currently for sale tells you almost nothing about what’s in the jar today. We look for recent testing — ideally within the last few months for flower — and we flag anything that feels like it’s been sitting on a shelf with stale documentation.

What we do when something doesn’t pass

We don’t carry it. That sounds simple and it is, but it’s worth saying explicitly because the decision point comes up more often than you’d expect. A supplier we like sends us a new batch and the COA shows a pesticide hit we didn’t expect. A terpene panel comes back flat on a cultivar that looked beautiful. A delta-9 number that’s technically compliant but closer to the line than we’re comfortable with.

Each of those situations gets a conversation with the supplier and a decision that prioritizes what goes in the jar over the relationship or the margin. That’s what a real COA process looks like — not checking a box, but using the document as an actual quality gate.

The COAs for every product we sell are linked directly on the product pages at sherlocksglass.com. If you’re ever curious about a specific batch, a specific test result, or how to interpret what you’re reading, our staff at any of our three locations in Raleigh, Durham, and Wake Forest can walk you through it. That’s part of what we’re there for.