A PDF arrives with a laboratory’s name at the top, two dozen numbers in the middle and a signature at the bottom. You have somewhere between an hour and a day to decide whether it justifies wiring a five-figure sum. Here is the uncomfortable part: every number on that certificate can be accurate, the lab can be real, the signature genuine — and the document can still cost you more than any defect, because of the basis on which one line is reported. The certificate is not lying. It is simply reporting a true number in the way that flatters the lot, and trusting that you will read it as the number you had in mind.
On a vanilla certificate of analysis, the line that most often costs buyers money is vanillin content — specifically whether it is reported on a dry basis or as received. The same lot can read 2.0 percent dry and about 1.3 percent as-is at 35 percent moisture. A supplier who quotes the dry figure and ships you the as-is reality has told you no lie and sold you less vanillin than you thought. Always confirm the basis, read moisture alongside vanillin, and require lot-specific results from an accredited lab.
This article is the reading order experienced buyers use, written down — and it starts with the vanillin basis because that is where the quiet money is. But a COA has six zones, and the others hide their own traps: a moisture figure that inflates the weight you pay for, a “complies” verdict with no value behind it, an accreditation scope that does not actually cover the test in question.
Zone One: The Line That Costs the Most
Vanillin is the headline number on a vanilla COA, and it is reported either on a dry basis — the vanillin as a fraction of the bean with all water removed — or as-is, on the bean including its moisture. The two can differ by a third or more. Consider a lot at 35 percent moisture reporting 2.0 percent vanillin on a dry basis. On an as-received basis, that same lot carries closer to 1.3 percent vanillin, because a third of the weight you are paying for is water, and water contains no vanillin.
If you are buying for extraction, vanillin is what you are actually paying for, and the basis is the difference between the yield you modelled and the yield you get. A supplier who quotes “2.0 percent vanillin” and ships a 35-percent-moisture lot has said something true and delivered something you did not price. The remedy is a single habit: never read a vanillin figure without reading the moisture figure and the stated basis in the same glance, and convert both to the same basis before you compare two lots.
| Lot | Vanillin (dry basis) | Moisture | Vanillin (as received) | What you’d actually pay per kg vanillin |
|---|---|---|---|---|
| A | 2.0% | 35% | ~1.3% | Higher than the dry figure suggests |
| B | 1.8% | 20% | ~1.44% | Lower per kg of vanillin — the “weaker” lot is stronger in practice |
Lot B looks weaker on the dry headline and is the better buy on an as-received basis, because it carries far less water. This single comparison, done on the right basis, routinely reverses which quote is cheaper — and it is invisible to anyone reading the headline vanillin figure alone.
A lot reported at 2.0 percent vanillin tells you nothing on its own. 2.0 percent on what basis, at what moisture? Until you have all three numbers, you have a marketing figure, not a measurement.
Zone Two: Is This Even a Real Test Result?
Before trusting any number, check who produced it. A certificate from a laboratory accredited to ISO/IEC 17025 for the specific test in question is a measurement. A certificate from a lab holding only ISO 9001 — a quality-management standard, not a competence-in-testing standard — or a “certificate” issued by the supplier themselves, is a declaration dressed as data. The accreditation also has a scope: a lab can be accredited for moisture and not for vanillin, and a result outside its scope carries none of the assurance the letterhead implies.
Accreditation and scope
Confirm ISO/IEC 17025 and that the specific parameter is within the lab’s accredited scope — not merely that the lab holds an accreditation for something.
Lot identity
The result must name the lot or batch you are buying, and that identifier must match the shipment. A generic COA reused across listings proves nothing about your beans.
Sampling
How the sample reaching the lab was drawn decides whether the result describes the lot or a selected handful. A perfect result on a sorted sample is a perfect result on nothing.
Values, not verdicts
“Complies” is not a result. Require the measured value and the specification it was tested against, so you can judge whether it passed comfortably or landed inside the measurement noise.
A lot reported at 1.82 percent against a 1.80 percent minimum has not comfortably passed — it has landed inside the measurement uncertainty, where the true value could sit on either side of the line. A verdict hides this; a value and its uncertainty reveal it. Always ask for the number.
Zones Three to Six: Physical, Safety, Provenance, and the Reply
With the vanillin basis understood and the lab verified, the rest of the certificate reads in a fixed order. Physical measurements — moisture, water activity, grade, count — confirm what you are paying for by weight. The safety panel — microbiology, heavy metals, pesticide residues, treatments such as ethylene oxide, and mycotoxins — matters most if you buy for extract, flavour or food manufacturing, where a residue failure can halt a production line. Provenance claims — origin, species, certifications — are only as good as the lot-level traceability behind them.
Moisture and water activity decide weight and shelf life; grade and count confirm you received the size band you bought. Read moisture next to vanillin, always.
Microbiology, heavy metals, residues and mycotoxins — check each against the limit for your market, and watch for a value reported on the wrong line or against the wrong standard.
Origin and species declarations and certifications are claims; lot identity and traceability are what make them checkable. A claim with no lot-level link is a hope.
A COA that comes up short needs a written response — missing basis, lead on the wrong line, verdicts instead of values — sent before you approve payment, not after.
The habit that ties it together is to read the certificate in the same order every time and to respond in writing to any gap before money moves. The six most common gaps — a missing vanillin basis, a heavy-metal figure on the wrong line, verdicts where values belong, missing marker compounds, a lot mismatch, and an unstated sample source — each have a standard reply that a competent supplier recognises and answers. The buyers who lose money are the ones who notice the gap after the wire, not before.
Even a flawless, accredited, lot-specific COA only describes the material the lab tested. Tie it to a representative sampling protocol and to contract clauses that let you test on arrival, or you are trusting that the tested sample and the container are the same thing.
The Six Gaps That Should Stop a Payment
Most problem certificates are not forgeries; they are honest documents with a gap in exactly the place a buyer tends not to look. Six recur often enough to be worth memorising, because each one, left unaddressed, quietly shifts risk onto you.
Missing vanillin basis
The headline vanillin figure with no “dry basis” or “as received” stated. Unusable until clarified, because the same number means two different lots.
Lead on the wrong line
A heavy-metal result reported against the wrong limit or in the wrong units. Easy to miss, and the difference between a pass and a market rejection.
Verdicts instead of values
“Complies” where a measured number belongs. Ask for the value and the specification it was tested against, every time.
Missing marker compounds
A certificate reporting vanillin alone, with no supporting marker profile, cannot distinguish natural vanilla from an addition. Require the markers if authenticity matters.
Lot mismatch
The lot identifier on the COA does not match the lot on the invoice or packing list. The certificate may be real and simply be for different beans.
Unstated sample source
Nothing on the document says how or by whom the tested sample was drawn. A perfect result on a sorted sample describes the sorting, not the lot.
Each of these has a standard written reply a competent supplier recognises and answers within a day or two. The discipline is to send the reply and withhold payment until it is answered — not to approve the wire and raise the question afterward, when your leverage is gone and the goods are already moving.
Turning Reading Into Terms
Reading a COA well is defensive; writing its requirements into the contract is offensive. The clauses that make certificates usable — requiring an accredited lab, a stated vanillin basis, lot-specific results, defined parameters with limits, and a remedy when a number fails — convert the certificate from a document you hope is honest into a standard the supplier has agreed to meet. How to Read a Vanilla Certificate of Analysis sets out all six zones in order, annotates a full certificate, and supplies the ten contract clauses and six ready-to-send replies that turn a lab PDF into something you can actually rely on before you pay.
Frequently Asked Questions
What is the most important line on a vanilla COA?
Vanillin content, read together with its reporting basis and the moisture figure. The same lot reads very differently on a dry basis versus as received — for example around 2.0 percent dry but nearer 1.3 percent as-is at 35 percent moisture — so a vanillin number without its basis and moisture is a marketing figure, not a measurement.
What does “dry basis” mean on a vanilla certificate of analysis?
Dry basis reports vanillin as a fraction of the bean with all water removed, which gives a higher number than the as-received basis that includes the bean’s moisture. Comparing a dry-basis figure from one supplier against an as-received figure from another makes the wetter, weaker lot look stronger, so always convert both to the same basis.
How do I know a vanilla COA is from a real laboratory?
Check that the lab is accredited to ISO/IEC 17025 and that the specific parameter is within its accredited scope — not merely that it holds some accreditation. A certificate issued by the supplier itself, or by a lab holding only ISO 9001, is a declaration rather than an accredited test result.
Why should I distrust a “complies” result on a COA?
Because a verdict hides the measured value and its uncertainty. A lot reported at 1.82 percent against a 1.80 percent minimum has landed inside the measurement noise rather than comfortably passing. Always require the actual value and the specification it was tested against instead of a pass/fail word.
Does a vanilla COA describe the whole shipment?
No. A certificate describes only the sample the lab tested. Unless it is tied to a representative sampling protocol and to contract clauses that let you test on arrival, a flawless COA can still sit alongside a shipment that does not match it.
Which safety parameters matter most on a vanilla COA?
For buyers making extract, flavour or food products, microbiology, heavy metals, pesticide residues, treatment residues such as ethylene oxide, and mycotoxins matter most, because a failure there can stop a production line. Each should be checked against the limit for your destination market and read as a value, not a verdict.
How do I respond to a vanilla COA that is missing information?
Send a written reply before approving payment that names the specific gap — a missing vanillin basis, a heavy-metal figure on the wrong line, verdicts instead of values, missing marker compounds, a lot mismatch, or an unstated sample source. A competent supplier recognises these requests and answers them; the buyers who lose money raise them after the wire, not before.
Further reading: ISO/IEC 17025 — testing and calibration laboratories