What a THCA Percentage Is Really Telling You (Less Than You Think)

The THCA percentage is the most-marketed number in this entire product category. Menus sort by it, prices scale with it, and a decimal point of difference gets treated like a verdict. So it's worth being precise about what the number actually is: the mass of THCA found in a small laboratory sample, divided by the mass of that sample, on one day, under one set of conditions. That's all. Useful — we build half our reviews on it — but carrying much less information than the marketing implies.

Here are the five layers between the number on the label and anything you'd actually experience, followed by how we audit the number itself.

Layer 1: The sample is not the bag

A lab tests a few grams drawn from a batch that may weigh kilograms. Cannabinoids are not distributed evenly across a plant — top colas that got full light typically run richer than lower buds from the same stalk. A conscientious producer samples representatively; an opportunistic one sends its prettiest colas to the lab and ships you the rest of the plant. Both get a COA. The number is honest about the sample; whether the sample was honest about the batch is a question the paperwork can't answer.

This is one reason batch-to-batch consistency tells you more than any single result — and why a strain that somehow tests identically high every batch is its own red flag (more on that in our reality check on high-THCA strains).

Layer 2: The moisture basis

Some results are reported dry-weight (moisture mathematically removed), some as-received. Flower is typically around 10% water, so the same material can print a meaningfully different number depending on basis — a 25% dry-weight result and a 22.5% as-received result can describe the same bud. Most COAs state the basis somewhere in the fine print. Most menus quoting the COA do not. When comparing two vendors' numbers, you may be comparing two accounting methods.

Layer 3: The 0.877 haircut

THCA is not THC. When heated, the molecule sheds mass as CO2, which is why every competent lab computes total THC = (THCA × 0.877) + delta-9. A bag labeled 28% THCA holds at most about 24.6% potential THC — the chemistry is worked through in decarboxylation, explained. Menus advertise the THCA figure rather than the total-THC figure for the obvious reason: it's the bigger number.

Layer 4: Delivery losses

Whatever the lab measured, you receive less. Combustion destroys a substantial share of cannabinoids in the ember; vaporizers extract more efficiently but still incompletely, and technique moves the needle in both cases (the tradeoff is quantified as well as public evidence allows in vaporizer vs combustion). The number in your lungs is the label number after three rounds of taxation: conversion loss, device loss, technique loss.

Layer 5: Everything the number doesn't measure

A cannabinoid percentage says nothing about terpene content — which is what flavor and aroma actually are — nothing about cure quality or smoothness, and nothing about freshness. A 30% bag with a fried, terpene-dead profile is a worse product than a fragrant 21% bag by every standard except the one on the shelf tag. If you're choosing between bags, the terpene panel is the more informative document: our terpene field guide covers how to read one.

The inflation problem

Now the uncomfortable layer under all five: the number is produced inside a bad incentive structure. Producers choose which lab to use. Higher numbers sell more flower. Labs that report higher numbers attract more producer business. State-regulated cannabis markets — with actual enforcement agencies — have repeatedly documented potency inflation and lab shopping. The hemp-derived THCA market has less oversight than those markets, not more. We are not accusing any specific lab of anything; we are observing that the referee is paid by one of the teams.

What we do about it, on every review:

• Recompute the math. Total THC must reconcile with THCA × 0.877 + delta-9. When we audited Fern Valley Farms, 25.582% THCA and 0.253% delta-9 reconciled to the printed 22.688% exactly — the system working as designed.
• Verify the lab exists and is accredited (ISO/IEC 17025, listed in the accreditor's directory), and that the report validates on the lab's own portal. Template and doctored COAs are a documented phenomenon: how to spot a fake COA.
• Apply a plausibility screen. Elite flower runs somewhere in the mid-to-high twenties total cannabinoids; the mid-thirties press against what trichome coverage can physically deliver. A menu full of 33% THCA flower is describing either historic agriculture or creative paperwork.

The heuristic that survives all of this

A verified 22% beats an unverified 32%. Every layer above cuts one direction: the headline number is a ceiling estimate produced under friendly conditions. So buy the documentation, not the digits — batch-matched COA from a verifiable lab, math that reconciles, recent test date, terpene panel published. That describes the vendors at the top of our 2026 vendor rankings, and it's point one through three of the checklist we audit with.

And keep the number in its lane at home, too: percentage is not a dosing instrument. Modern flower is strong across the board, individual response varies widely, and the difference between 22% and 28% is smaller than the difference between one draw and three. Start low, go slow — the honest dosing guide — and remember the number the label never prints: heated THCA is THC, in your body and on a drug test.

The percentage is one instrument reading from one moment. Treat it like a data point and it serves you; treat it like a verdict and it sells you.