What actually decides how chocolate tastes

Popular writing about chocolate flavour starts with origin — Madagascan citrus, Ecuadorian floral, Ghanaian classic — because origin is the part that is printed on the wrapper, sounds like wine, and is easy to write about.

It is also, on the available evidence, close to the weakest of the causes. Growing conditions matter, and they matter less than what happened to the beans afterwards and much less than what the maker put in the bar.

The list below is a rough ranking rather than a measured one. Interactions are real, a catastrophic failure at any stage dominates everything above it, and the ranking describes typical bars rather than every bar. It is offered because having roughly the right order is far more useful than having no order at all, and the order most readers arrive with is close to reversed.

A 70% dark bar and a 32% milk bar made from the same beans by the same maker taste almost nothing alike, and no fermentation or roasting difference produces a gap that large.

SUGAR is the biggest single lever. It supplies sweetness, dilutes everything else, and suppresses the perception of bitterness and astringency. Moving from 70% to 85% cocoa is mostly a change in how much sugar is present.

MILK SOLIDS add sweetness, dairy flavour, and a coating fattiness that mutes cocoa character and acidity. They are why milk chocolate reads as mild regardless of what the beans were.

ADDED COCOA BUTTER raises the fat fraction without adding cocoa solids, which softens the melt and dilutes flavour intensity — so two bars at the same stated percentage can differ substantially depending on how that percentage splits between solids and butter. The percentage does not tell you the split.

INCLUSIONS AND FLAVOURINGS override nearly everything. A bar with salt, or nuts, or fruit, or vanilla is a bar whose dominant flavour decision was made in the recipe.

The largest bean-side cause, and the one this catalogue returns to repeatedly. Fermentation generates the precursor compounds that roasting later develops, and it is where most of the character a reader would call 'origin' actually comes from.

It is also where the largest faults originate. An under-fermented bean produces flat, astringent chocolate; an over-fermented one produces harsh or putrid notes. Neither is recoverable downstream, which is what puts fermentation above roast: a roaster can shape what fermentation produced and cannot supply what it failed to produce.

Roasting develops the precursors into aroma compounds through the Maillard reaction and drives off some volatile acidity. It is the maker's largest single expressive decision and it operates within limits fermentation set.

Light roasting preserves more delicate and fruity character and risks leaving astringency and raw notes. Dark roasting builds toasted, coffee-like depth and flattens origin distinctiveness. Both are legitimate choices rather than a right and a wrong answer, and a maker's house roast style is often more identifiable across their range than any origin is.

The variety sets the range of what is possible. Some material carries a genuine potential for fruit, floral or nut character; some does not, and no processing will conjure it.

It sits fourth rather than first because potential is not expression. Well-regarded genetics badly fermented produce poor chocolate, and ordinary bulk material handled carefully produces good chocolate. The catalogue's variety records are explicit that the popular four-group account is a trade classification rather than a genetic one, which is a further reason not to lean on it.

Soil, altitude, rainfall and shade influence the tree and the pod, and a claim about a named valley can be checkable in a way a claim about a country cannot.

The reason it ranks here rather than higher is scale. A country contains hundreds of thousands of farms with different genetics, different fermentation practice and different post-harvest infrastructure, so a country-level origin claim is averaging across more variation than it is describing. Regional and estate-level claims carry more information for exactly that reason.

These shape texture and moderate flavour rather than creating it. Conching drives off residual volatile acidity and rounds harshness; refining decides whether the bar reads as smooth or gritty; tempering and cooling decide gloss, snap and how the chocolate melts, which changes the rate at which aroma is released.

A fault here is very noticeable — a gritty or badly tempered bar is unpleasant regardless of the beans — but within the normal range these steps adjust rather than determine.

Last, and capable of overriding everything above it. Chocolate stored warm, damp, or next to anything aromatic picks up what it was stored beside and loses what it had. Milk and white chocolate stale faster than dark.

This is worth ending on because it is the only item on the list a reader controls. Everything from genetics to conching was decided before the bar was bought; how it is kept afterwards is the one variable still in play, and a well-made bar can be spoiled entirely by a warm cupboard.