Fermentation, explained at the tank

A coffee cherry is fruit, and the seed we roast sits inside pulp, a layer of sticky sugary mucilage and a parchment shell. Processing is the work of getting the seed out and dry without letting the fruit spoil it. Fermentation is what happens while that is going on, whether anyone plans it or not, and it accounts for a great deal of what you taste later.

The mucilage is roughly water, pectins, sugars and acids. Left alone, yeasts and lactic acid bacteria that live on the fruit and on the equipment begin to eat those sugars. Yeasts produce alcohols, esters and carbon dioxide; lactic bacteria produce lactic acid and a soft, dairy-like roundness; acetic bacteria, given oxygen, produce acetic acid, which in small amounts reads as bright and complex and in large amounts as vinegar. None of these compounds are inside the seed at the start. They migrate in, along with their aromatic by-products, while the pectin structure breaks down.

Four routes, one seed

• Washed: pulp removed, then the mucilage fermented off in tanks for a short window and rinsed clean before drying. Fermentation is a tool for removing sugar, not for adding flavour, so the cup shows the seed itself — clear acidity, defined structure, less fruit noise.
• Natural: the whole cherry dried intact, often on raised beds, for two to four weeks. The fruit ferments slowly around the seed. Heavier body, lower perceived acidity, strong berry and stone-fruit character, and a much higher risk of mould or defect if the drying is not turned and watched.
• Honey: pulped but dried with some or all of the mucilage still attached — yellow, red or black depending on how much is left and how much shade is used. It sits between the two: fruit and syrupy sweetness with more clarity than a natural.
• Controlled anaerobic: cherries or de-pulped seeds sealed in tanks without oxygen, often with pressure release and controlled temperature. Acetic bacteria are suppressed, lactic and yeast activity favoured, and the producer can steer the result with time, temperature and pH rather than hope.

Why one lot sings

Two lots of the same variety from the same hillside, picked in the same week, can taste of apricot and of very little. The variables are not mysterious, they are just numerous. Ripeness at picking sets the sugar available to the microbes; a batch with fifteen per cent underripe cherries ferments differently and tastes thinner. Ambient temperature changes which organisms dominate — warm nights favour fast, boozy activity, cool ones a slower and cleaner build. Tank hygiene decides whether the population is broadly the one the producer wants. And drying is where most good fermentations are lost: dried too fast and the seed cracks and stales early, too slowly and the fruit keeps fermenting past the point of pleasure.

What we buy, what we pass

We taste blind, we taste the same lot twice at least a fortnight apart, and we buy for sweetness and clarity first. A processed-forward coffee has to earn its place: if the fruit is loud but there is no structure underneath it, that is a lot that will be tiresome by the second cup and unrecognisable in three months.

What we avoid is easier to list. Boozy, solvent-like alcohol that sits on top of everything else. Sharp acetic bite. That cider or overripe-bin note that tells you the tank ran too long. And, above all, instability: lots that arrive extraordinary and are dull or vinegary by the third month, because a coffee we cannot serve consistently for a season is no use to six cafés and a weekly roast schedule.

Fermentation is not a flavour you add. It is a controlled loss of sugar, and the skill is knowing when to stop.