AarleezWellness
Laboratory analysis of a botanical extract sample

Insights — Quality & compliance

How to read a Certificate of Analysis

A COA is the only thing standing between a specification and a hope. Most buyers glance at the assay figure and file it. Here is what the rest of the document is telling you.

A Certificate of Analysis is a laboratory's statement about a specific quantity of material. That last part is where most of the value sits, and where most of the problems hide. A COA that cannot be tied to the drum in front of you is a brochure with numbers on it.

Start at the batch number, not the assay

The first thing to check is not the headline percentage. It is whether the batch or lot number on the certificate matches the batch number printed on the container you received. If it does not, nothing else on the document is evidence of anything. This sounds obvious and it is the single most common failure we see when buyers bring us an incumbent supplier's paperwork to compare.

Alongside it, look for a manufacturing date and a retest or expiry date. A certificate with no dates is not traceable, and a certificate whose manufacturing date is two years before your delivery tells you something about how long the material has been sitting.

The assay figure is meaningless without a method

"5% withanolides" is a number. "5% withanolides by HPLC" is a specification. The difference is not pedantry — it is often a difference of several-fold in what is actually present.

Gravimetric and UV-spectrophotometric methods measure broad classes of compounds and will read anything with roughly the right behaviour. High-performance liquid chromatography separates and quantifies individual compounds. For ashwagandha, a gravimetric "withanolide" figure can be several times higher than the HPLC figure on the same material. Both certificates are technically honest. Only one of them tells you what you are dosing.

So: for any extract standardised on a marker compound, look for the method next to the number. If it is absent, ask. A supplier who cannot tell you the assay method within a working day is not close enough to the material.

Plant part and extraction medium

These two fields do more work than their size on the page suggests. Ashwagandha root and ashwagandha leaf are different materials with different compositions and very different costs. Turmeric rhizome and turmeric leaf are not interchangeable. If the plant part is not stated, the safe assumption is that stating it would not have helped the sale.

Extraction medium — water, hydro-alcoholic, alcohol — determines which compounds came across and which stayed behind. It also determines whether a residual solvent test is relevant, which brings us to the next section.

The safety panel: what should always be there

Beyond identity and assay, a complete certificate covers four safety areas. If any of them is missing, ask why before you ask about price.

  • Heavy metals. Lead, arsenic, cadmium and mercury as a minimum. Botanicals concentrate what is in the soil, and a 10:1 extract concentrates the heavy metals ten-fold along with everything else. This is one of the few places where a concentrated extract carries more risk than the raw herb.
  • Microbiology. Total plate count, yeast and mould, and the pathogens: Salmonella, E. coli, and usually Staphylococcus aureus. If your finished product has no kill step, these limits are the whole ball game.
  • Pesticide residue. A multi-residue screen, not a statement that the material is "pesticide free", which is not a testable claim.
  • Residual solvent. Only relevant where solvent extraction was used, which is exactly why the extraction medium field matters.

Six things that should make you ask a question

  • Round numbers everywhere. Real laboratory results are untidy. An assay of exactly 5.00%, moisture of exactly 5.0% and a plate count of exactly 1,000 across successive batches is a document that was typed rather than measured.
  • Identical figures across batches. Botanicals are agricultural. If three consecutive lots report the same numbers to two decimal places, the certificate is a template.
  • "Complies" instead of a result. A limit with the word "complies" or "conforms" next to it is an assertion. A limit with a measured value next to it is a result. You want results.
  • No laboratory identified. In-house testing is normal and acceptable, but the certificate should say whose laboratory it was. An unattributed report cannot be followed up.
  • A specification range instead of a batch value. "Withanolides: 5–7%" is what the supplier can produce. "Withanolides: 5.4%" is what you bought.
  • Poor reproduction. A certificate that has been scanned, photographed and re-scanned has passed through hands. Ask for the original from the source.

What to do with all this

Two practical habits protect you more than any amount of document analysis. First, ask for the certificate before you order, tied to the specific batch being quoted, along with a sample drawn from it — then have your own laboratory verify the assay independently on at least the first delivery. Second, keep the certificates. When a question surfaces eight months later about a batch you bought in February, the paperwork is what turns an argument into an investigation.

We send a batch-tied Certificate of Analysis with every consignment, and we hold retained samples against every lot, precisely so that later question has something to test. If you would like us to look over a certificate from your current supplier, send it across with the name redacted — we will tell you honestly what it does and does not establish.

Next step

Have a specification in mind?

Send us the line, the specification and the volume. You will get the certificate for the real batch and a delivered UK price.