Because knowing the origin, the price or the supplier measurably changes what you perceive. Coding the samples removes an influence you cannot otherwise control for.

Buyer’s Guide

Green Coffee Sample Evaluation

Sample evaluation is where a purchase decision is actually made. This page covers the method, and the record-keeping that makes the method worth running.

Last reviewed August 20263 min read

Quick answer

Evaluate a sample in three passes: a physical assessment of the green, a sample roast, and a blind cupping in multiple cups against a reference you already buy. Record everything against the lot identity — an evaluation you cannot reproduce six months later has limited value.

Know which sample you have

Sample types and their standing
Type	What it is	Standing in a contract
Offer sample	Sent to support an offer, drawn from the parcel	Indicative — the basis for shortlisting
Type sample	A reference representing a quality level rather than a specific parcel	The parcel must be equal to or better than it
Approved / retained sample	The sample the buyer approved, sealed and retained by both parties	The reference the contract is enforced against
Pre-shipment sample	Drawn from the milled parcel before stuffing	Verifies the actual shipment against the approval
Arrival sample	Drawn at destination	Basis for any arrival claim

Ask, do not assume

Approving an offer sample and treating it as an approved sample without sealing and retaining it leaves nothing to enforce against later. If the sample matters, seal it and log it.

Pass one: the green

Colour and uniformity — blue-green through to pale; mottling suggests uneven drying.
Odour — should be neutral to faintly sweet. Ferment, mould or chemical notes in the green are disqualifying.
Moisture and water activity, on your own instruments, recorded with the method.
Screen analysis against the offered specification.
Defect count on a defined sample, against a stated weighting table.
Bulk density, if it informs your roast planning.

Pass two: the sample roast

Sample roasting is where the evaluation either becomes comparable or does not. Use a consistent profile to a defined colour, rest for a consistent period, and roast all samples in a comparison session the same way.

Record the profile and the resulting colour measurement, not just "light".
Count quakers after roasting — they are invisible in the green and they tell you about picking discipline.
Note anything that behaves unexpectedly: uneven development points at a wide size distribution or uneven moisture.

Pass three: cupping

Follow a consistent protocol — see coffee cupping — and observe three rules that matter more than any scoring detail.

1Multiple cups per sample. Faults occur at bean level. One cup will miss them.
2Blind. Code the samples. Knowing the price or the origin changes what you perceive.
3Against a reference. Include something you already buy and know well. Absolute judgements drift; relative ones do not.

Recording

The purpose of a record is to let you answer, next season, why you bought or rejected something. Keep it against the lot identity, not against the supplier name.

What to log

Lot identity: Origin, station or factory, outturn, crop year
Sample type and date: And who sent it
Physical results: Moisture, aw, screen, defects, density — with methods
Roast record: Profile, colour, quaker count
Cupping notes: Per cup, plus any faults, plus the score if you use one
Decision and reason: Bought, rejected, or held — and why

When a sample disappoints

Say what was wrong, specifically. "Not for us" tells the exporter nothing. "Phenolic in two of five cups" is actionable.
Ask whether it is representative. A single poor sample can be a sampling problem rather than a lot problem.
Ask what else is available against the same specification — a rejection is a data point for the next offer, not the end of the conversation.
Retain the sample. If a similar problem recurs, the earlier sample is evidence.

Enquiries

Tell us the specification and we will come back with the lot identity and the physical analysis stated.

Request Current Crop Offer

Frequently asked questions

How much coffee is needed for a proper sample evaluation?

Enough to run a physical analysis, a sample roast and a cupping with multiple cups, with some left over to retain. In practice a few hundred grams is the working minimum for a full evaluation, and more is better if the parcel is large.

What is a type sample?

A reference representing a quality level rather than a specific parcel. A contract on type sample terms requires the shipment to be equal to or better than the type, which is a weaker basis than an approved sample drawn from the actual parcel.

Should I count quakers in the green?

You cannot — that is the point of them. Quakers are under-developed beans that look normal in the green and fail to brown in the roast. Count them after sample roasting.

Tell us the coffee you need

Lots can be specified by origin, region, process, grade, screen, moisture, defect tolerance, crop year and packaging. Send what you know and we will confirm what each origin realistically supports.