Cut Test 101: The ICCO 100-Bean Method for Grading Cacao
A step-by-step guide to the cut test — the single most useful quality-control tool a bean-to-bar maker has. Covers the ICCO 100-bean method, the six grading categories, how to interpret results for fine-flavor vs commercial cacao, when to accept a lot, when to ask for a discount, and when to walk away.
The cut test is the single most informative fifteen minutes a bean-to-bar maker will ever spend on a new lot of cacao. You slice 100 beans in half, look at the interiors, and count six things. When you're done you know more about how well your cacao was fermented than any supplier's marketing copy will ever tell you. This post is the step-by-step guide — the ICCO standard, the categories, the targets, and what to do with the results.
What the cut test actually tells you
The cut test — formally called the ICCO 100-bean cut test — is the traditional method used across the cacao industry to grade a sample for fermentation quality, drying quality, and major defects. The underlying principle is that the interior color and texture of a dried cacao bean is a visible record of what happened during fermentation. Well-fermented beans develop a characteristic brown interior; under-fermented beans retain purple-to-slate coloring; moldy or insect-damaged beans show specific, recognizable patterns.
The test is cheap, fast, and diagnostic. It won't tell you exactly how the finished chocolate will taste, but it will tell you whether the lot has been fermented well enough to produce fine-flavor chocolate at all — which is the first question a serious maker needs answered before roasting anything.
When to run a cut test
- On a sample before purchase. The cut test is the single most useful step in evaluating a new supplier or a new lot. Before you commit to 25–200 kg of cacao, slice 100 beans.
- On arrival of a purchased lot. Verify that what you received matches what you sampled. Cacao can shift in transit (especially moisture) and unscrupulous suppliers occasionally swap lots between sample and shipment. Independent verification takes 15 minutes.
- Before every production batch. Most mature makers sample each production batch's incoming beans with a brief cut test — a 25- or 50-bean version — as a final QC step. Catches storage-related issues before they reach the roaster.
- As a baseline documentation record. Keep your cut-test results on file per lot. When a finished bar tastes different from its previous version, the first thing to check is whether the cut-test grade changed.
Equipment
The cut test is gratifyingly simple. You need:
- A guillotine cutter (purpose- built for cacao; ~$60–$200 from specialty suppliers) or a sharp, sturdy knife with a cutting board;
- A 100-bean tray — a shallow board with 100 slots, ideal but optional; any flat surface works;
- A digital scale accurate to 0.1g for bean-weight statistics;
- A notebook or spreadsheet to record results;
- Decent natural light or a good bench lamp.
A dedicated cacao guillotine is a worthwhile purchase once you're running cut tests monthly — the blade is sized for the bean, cuts cleanly in half, and doesn't crush. Most small makers start with a kitchen knife and upgrade within a year.
Step-by-step protocol
- Draw a representative sample. From the top, middle, and bottom of the sack, scoop a small handful each. Combine and mix. Pick 100 beans at random — don't select for appearance. If you cherry-pick, you're testing your eye, not the lot.
- Weigh the 100 beans. Divide by 100 for average bean weight. Specialty cacao usually falls between 1.0g and 1.4g per bean. Consistent weight suggests consistent bean size, which is good for roast uniformity.
- Cut each bean lengthwise. Hold the bean on its edge, slice from tip to tip. A guillotine makes this one motion. You want the interior halves to lie flat so you can see cotyledon color clearly.
- Grade each bean into one of the six categories below.
- Tally the counts. Record the percentage of each category on your lot record.
- Compare against targetsfor the class of chocolate you're trying to make.
Total time: 12–20 minutes for the full 100-bean test once you're practiced. First time you do it, allow 30–40 minutes — the grading decisions slow down until your eye adjusts.
The six grading categories
Every bean falls into one of these six categories. Your job is to recognize them, not to agonize over edge cases — if a bean is ambiguous, pick the worse category to be conservative.
| Category | Visual | What it means |
|---|---|---|
| Fully brown | Uniform brown cotyledon; visible natural cracks; dry, chalky texture | Well-fermented, properly dried — the target |
| Partly brown / partly violet | Mixed brown-and-purple interior; some separation between cotyledon halves | Under-fermented; will produce a slightly astringent, green bar |
| Fully violet / slate | Solid purple-to-gray interior with no brown; often compact, waxy texture | Significantly under-fermented; harsh, sharp flavor; major defect |
| Moldy | Whitish or greenish fuzz visible on or in the cotyledon; musty smell | Storage or fermentation failure; reject |
| Germinated | Visible sprout or embryo; cotyledon often split or deformed | Drying failure; typically produces off-notes |
| Insect-damaged / hollow | Visible tunneling, frass, empty sections, or cracks inconsistent with drying | Handling or storage failure; reject |
Target percentages by chocolate class
Different chocolate classes have different acceptable cut-test profiles:
| Class | Fully brown | Partly brown | Slate / violet | Defects |
|---|---|---|---|---|
| Fine-flavor specialty (elite) | 70%+ | <25% | <5% | 0% |
| Fine-flavor (standard) | 60%+ | <30% | <10% | <3% |
| Commercial / wholesale | 40%+ | <45% | <20% | <5% |
| Bulk / industrial | >25% | anything | anything | <10% |
How to interpret and respond
Once you have cut-test numbers for a lot, the question becomes what to do with them. The decision tree most specialty buyers use:
- Exceeds target for intended class. Proceed to purchase or production. Keep the result on file as baseline for future comparison.
- Misses fine-flavor target but passes commercial. Options: negotiate a lower price that reflects the true class; use the lot for blends rather than single origin; reserve it for bars where the origin character is less critical. Don't pay a fine-flavor price for commercial cacao.
- Misses commercial target.The lot is either bulk-quality or defective. If it's already yours, explore blending it into inclusion bars or selling it at a discount to another maker. If you're evaluating pre-purchase, walk away.
- Defects over 2%. Reject or return. No price negotiation will produce good bars from a defect-heavy lot.
The supplier said 68% fully brown. My cut test said 41%. I showed him the photos. He sent me a replacement lot the next week. He didn't argue because the numbers don't lie. The test paid for itself the first time I ran it.
Common mistakes
- Cherry-picking beans.If you select 100 “typical-looking” beans by eye, you're biased high. Grab a mixed handful and cut whatever you grab.
- Not sampling from multiple depths. Cacao beans settle during transit; the top and bottom of a sack can have different moisture and different fermentation grades because suppliers sometimes layer. Always sample from top, middle, and bottom.
- Grading in bad lighting. The difference between partly-brown and slate is subtle. Natural daylight or a neutral-temperature bench lamp matters.
- Being too generous. When in doubt, grade down. Consistent, disciplined grading is more valuable than optimistic grading.
- Skipping the record. A cut test without documentation is a moment in time; a cut test saved per lot is institutional memory. Write the numbers down every time.
Common questions
How often should I run a full 100-bean test?
On every new lot purchase, without exception. On every production batch, a quicker 25- or 50-bean test is often sufficient as long as the full test on the lot was clean.
What if my results disagree with the supplier's CoA?
Share your results with the supplier and ask them to re-verify. Minor variation (±5% fully brown) is normal due to sampling differences. Major disagreements (±15% or more) suggest either your sampling technique is off or the supplier's grading is loose. In either case, photograph your test and document it — that documentation is your leverage in any commercial dispute.
Can the cut test predict finished-chocolate flavor?
Not precisely, but it predicts the ceiling. A lot with 75% fully brown has the genetic and fermentation potential for excellent chocolate if roasted and processed well. A lot with 35% fully brown has a ceiling — no amount of good roasting will recover the under-fermented beans' potential.
Is there a way to partially rescue an under-fermented lot?
Extended longer fermentation is not realistic post- purchase (the beans are already dry). Slightly more aggressive roasting and longer conching can mask some of the astringency, but the missing flavor precursors are gone — you can't generate them in the finished bar later. Under-fermented cacao is what it is. See our fermentation guide for the underlying chemistry.
The cheat sheet
| Question | Short answer |
|---|---|
| Time required? | 15 minutes once you're practiced |
| Equipment? | Guillotine or knife, scale, notebook |
| Fine-flavor target? | 70%+ fully brown; <5% slate; 0% defects |
| Commercial target? | 40%+ fully brown; <20% slate; <5% defects |
| When to reject? | Any defects over 2%; or missing commercial target entirely |
| Most important habit? | Sample from top, middle, and bottom; grade down when in doubt |
The cut test is one of the very few craft chocolate practices where a serious maker and a casual maker are distinguishable purely by discipline. The test itself is simple — anyone can learn it in an hour. The makers who actually run it on every lot, record the results, and use the numbers to make purchasing and production decisions end up with a quality consistency their peers can't match. The makers who skip the test “just this once” end up with the bar that tastes different this month and no idea why.
For the chemistry of what a good cut test actually represents, read our fermentation primer. For how cut-test results plug into the broader sourcing conversation, see our direct-trade sourcing guide. And for how to translate cut-test quality into an origin-appropriate roast, our roast profile design guide closes the loop.