Criollo vs Forastero vs Trinitario: How to Choose Cocoa Beans for Chocolate Production

The Variety Name Is Only the First Filter

Criollo, Forastero and Trinitario are still useful words in trade conversations. They tell you what kind of cocoa the supplier claims to offer and what flavour direction may be expected.

They do not tell you whether the lot is usable.

Modern cacao genetics has moved beyond the old three-type model. Research has identified multiple genetic clusters, while the cocoa trade often works with a more practical distinction: fine or flavour cocoa versus bulk cocoa.

Even that split has exceptions. Some Forastero-type cocoas can be fine flavour; some Trinitario-type cocoas are traded as bulk.

A 72% single-origin bar, a milk shell, a hazelnut filling and a bakery chocolate need different cocoa decisions.

Quick Working Comparison

Criollo: Do Not Pay for the Name Until the Lot Passes the Bench

Criollo-type cocoa can be excellent in a fine dark bar. A clean lot may show floral notes, honey, nuts, dried fruit or soft red fruit with less aggressive bitterness than many bulk lots.

That is useful only if the lot is clean.

For a premium bar, reject a Criollo-type lot with visible mould, strong smoke odour, heavy slaty beans or harsh acidity that stays in the liquor after roasting. The defect will not become elegant because the label says Criollo.

The roast should also be restrained. If the buyer pays for delicate aroma, a heavy roast can remove the value. A trial should check the same lot at least in nibs, liquor and finished chocolate.

Criollo-type cocoa is usually justified when:

• the product is a premium dark bar or tasting-led line;
• the lot volume is small enough to manage separately;
• the supplier can explain origin, harvest and post-harvest handling;
• the finished chocolate sells the aroma, not only the origin story.

It is usually weak economics for bakery, strongly flavoured fillings or very sweet milk recipes.

Forastero: Useful Cocoa, But Bad Lots Should Still Be Rejected

Forastero-type cocoa is often treated as the industrial option. That is not a problem by itself. A clean Forastero-type lot can be exactly right for chocolate that needs cocoa body, stable supply and predictable cost.

The problem starts when “industrial” becomes an excuse for poor raw material.

Mould, smoke contamination, insect damage and heavy slaty content are not solved by conching. They become sorting loss, flavour correction, extra cocoa butter, extra sugar, complaints or rejected batches.

Forastero-type cocoa often works well for:

• bakery chocolate;
• fillings;
• coatings;
• compound-style applications where allowed;
• milk chocolate with strong cocoa body;
• blends where consistency matters more than single-origin nuance.

Incoming control should be strict: odour, moisture, cut test, bean count, shell percentage, test roast and liquor tasting. If the lot is cheap but gives poor nib yield or high shell carryover after winnowing, the real cost is not cheap.

Trinitario: Good Starting Point, Still Too Broad to Buy Blind

Trinitario-type cocoa is often the most practical premium category for a growing chocolate factory. It can give fruit, nuts, floral notes or rounded cocoa body with better availability than rare Criollo-type lots.

But Trinitario is a broad hybrid category. Two lots can behave completely differently.

One may become clean and expressive after a medium roast. Another may stay acidic, thick or flat after conching. A third may be sold as premium but behave like ordinary bulk cocoa in the finished bar.

For Trinitario, tasting raw nibs is not enough. Run the lot through a small version of the real process:

• roast;
• crack and winnow;
• measure nib yield;
• refine;
• conche;
• check viscosity;
• temper;
• mould;
• taste after crystallisation and again after several days.

If the lot loses aroma after conching or becomes difficult to process at the target recipe, the category name no longer matters.

Moisture: Specify the Number and the Method

For fermented and dried cocoa beans, a practical target is often around 6.5–7.5% moisture. Around 7.5% is widely treated as suitable for storage; near or above 8%, mould and rehumidification risk become more serious.

Do not write only “max 8%” in a contract. Specify:

• target range;
• maximum allowed value;
• measurement method;
• sampling method;
• when the measurement is taken;
• what happens if bags arrive above specification.

For cocoa, use a cocoa-relevant standard or agreed laboratory oven method. If a handheld moisture meter is used at intake, it should be calibrated against the agreed reference method.

Practical intake check: measure several bags, not only the cleanest sample. Moisture can vary inside a shipment, especially after humid storage or container transport.

If moisture is too high, mould and OTA risk become more important. If it is too low, beans can become brittle, create more fines and make winnowing less clean.

Cut Test: Put Defect Thresholds in the Specification

The cut test should classify the beans, not just confirm that “they look fine.”

At minimum, record:

• well-fermented beans;
• slaty beans;
• purple or under-fermented beans;
• mouldy beans;
• insect-damaged beans;
• germinated or flat beans;
• smoky or abnormal odour.

For fine chocolate, tolerance should be tighter than for bulk industrial use. A useful purchasing specification should define maximum allowed percentages for mould, slaty beans and insect-damaged beans. Generic “good quality” language is not enough.

Mould is critical for almost any product. Heavy slaty content is especially damaging for fine dark bars because it often brings weak fermentation, bitterness and poor flavour development. Smoke is a serious problem for delicate chocolate; phenolic notes fr om fire drying are difficult to remove later.

Add photo examples to supplier agreements when possible. “Slaty” and “purple” are not always interpreted the same way by every seller.

Mycotoxins and Food Safety: Do Not Treat Mould as Only a Flavour Problem

Mould is not only a sensory defect. It can indicate increased risk of mycotoxins, including ochratoxin A and, depending on origin and storage conditions, aflatoxin concerns.

The practical rule for B2B buying is simple: if moisture, storage or origin risk is high, request a Certificate of Analysis that includes relevant microbiology and mycotoxin testing according to the target market and customer requirements.

The contract should define:

• which mycotoxins are tested;
• analytical method;
• maximum accepted limits;
• who pays for retesting;
• what happens if the lot fails.

Do not rely on “looks dry” or “smells acceptable” for high-risk shipments.

Shell Percentage and Real Cost

Price per tonne is not the same as cost per usable kilogram.

The shell is normally a significant non-edible fraction of the bean. Higher shell means less edible nib. It can also mean more winnowing difficulty. If shell fragments remain in the nib fraction, bitterness, astringency and gritty perception can increase.

Calculate cocoa cost like this:

real nib cost = bean price / usable nib yield

Then adjust again for:

• sorting loss;
• roasting loss;
• shell carryover;
• nib loss during winnowing;
• rejected material;
• extra cocoa butter needed for viscosity correction.

For a serious purchase, test 25–50 kg if possible. A tiny sample can show flavour, but it will not give reliable data on winnowing, shell percentage and process loss.

Particle Size: Use D90, Not Only an Average

“18 microns” sounds precise, but it is often incomplete.

For chocolate texture, the larger particles matter because they are the ones the mouth detects. Laser diffraction results are usually read through values such as D50 and D90. D90 means 90% of particles are below that size.

For a smooth premium bar, a useful target may be D90 below about 25 µm, depending on recipe and measurement method. For fillings or bakery applications, a slightly coarser profile may be acceptable if the product does not feel gritty and the process runs better.

Over-refining creates another problem. Too many fines increase surface area, so the mass can become thick and need more cocoa butter or emulsifier. That affects cost, dosing, enrobing, moulding and pumpability.

A production specification should include:

• particle size method;
• D90 target;
• acceptable percentage above 30 µm, if used;
• viscosity or yield value at a defined temperature;
• product application: bar, filling, coating or bakery.

Do not set particle size without checking viscosity.

Viscosity After Refining and Conching

Bean choice affects viscosity indirectly through fat content, particle shape, fibre, shell carryover and the way solids break down during refining.

For production, the useful measurement is not “the chocolate feels fluid.” Define the test:

• temperature, often 40°C for molten chocolate rheology;
• instrument or viscometer method;
• shear conditions;
• viscosity range;
• yield value, if relevant;
• measurement point: after refining, after conching, after holding.

A liquor that tastes good but creates excessive viscosity may require more cocoa butter, different lecithin dosing, longer conching or a revised particle size target. That changes cost and equipment load.

For a new cocoa lot, measure viscosity after refining and again after a defined conching time. If the mass is already too thick after refining, conching will not magically make the economics better.

Roasting: Record the Profile, Not Just the Result

A pilot roast should be documented.

Minimum useful data:

• batch size;
• charge temperature;
• time-temperature curve;
• end temperature;
• roast loss by weight;
• colour change, if measured;
• sensory notes for nibs and liquor.

For delicate Criollo-type cocoa, the question is whether the aroma survives. For harsher bulk cocoa, the question is whether acidity and bitterness soften enough without creating burnt notes.

A supplier sample that works only under one fragile roast profile may be risky for larger production. Repeatability matters.

Choosing by Product Type

Fine Dark Bar

Use Criollo-type or high-quality Trinitario-type cocoa only if the cut test is clean and the finished chocolate proves the aroma.

Reject mould, strong smoke, heavy slaty content and harsh acidity that stays after roasting. These defects are too visible in a high-cocoa bar.

Milk Chocolate

Milk can soften acidity and bitterness, but it also covers delicate notes. A rare Criollo-type lot may not pay back in a sweet dairy-forward recipe.

A rounded Trinitario-type cocoa or clean Forastero-type blend may give better cocoa body, viscosity and cost.

Fillings and Pralines

For fillings, origin character is often less important than compatibility with nuts, cream, fruit, alcohol or spices.

Watch viscosity closely. If fine refining makes the filling base too thick, the issue becomes dosing, texture and shelf stability, not bean prestige.

Bakery and Heat-Treated Products

Heat can erase delicate aroma. For bakery, test the baked product, not only the liquor.

Forastero-type cocoa or blends often make more sense here because the product needs colour, cocoa strength and cost control.

Premium Bean-to-Bar Range

Trinitario-type cocoa is often a practical starting point. It gives room for origin differentiation without relying entirely on scarce micro-lots.

Still, each origin needs its own roast and conch decision. A single standard profile may make production easier but flatten the differences customers are paying for.