Where the taste actually comes from
Flavour is decided before you grind.
The extraction chapter is about not ruining a coffee. This one is about what there was to ruin. By the time a bag reaches your grinder, four decisions have already fixed most of what it is capable of tasting like — and you made none of them.
Four decisions, in order
A coffee's flavour is not one thing that happens; it is a sequence of narrowings. Each step can only work with what the step before it handed over, which is why the last one — roasting — is both the cheapest to change and the most destructive.
| Decision | Made by | What it fixes |
|---|---|---|
| Species | The plant | The ceiling. Sugars, acids and caffeine are set by which plant this is |
| Origin & elevation | Geography | How slowly the cherry ripened, and therefore how much is in it |
| Process | The farm | How much of the fruit ends up in the seed |
| Roast | The roaster | How much of the three above you are still allowed to taste |
Two plants, not two grades
Almost all coffee is one of two species, and the difference between them is not quality — it is botany. Arabica is a fussy highland plant: it wants cool nights, it wants altitude, and it catches every disease going. Robusta is a tough lowland one that shrugs off heat and leaf rust and yields more per tree.
They pay for those habits in the cup. Arabica carries roughly twice the sugars and lipids, which is where sweetness and aroma come from. Robusta carries roughly twice the caffeine — about 2% against 1% — and caffeine is bitter. The deck puts it less diplomatically: refined against burnt rubber.
That is not a reason to sneer at it. Robusta is what puts a thick, stable crema on a traditional Italian espresso, and it grows in places where an arabica plant would simply die.
The belt
Coffee grows in one band around the planet: roughly between the two tropics, where the average sits around 21–29 °C and it never frosts. Everything you will ever drink came from inside that strip. It is why "which origin" and "which climate" are the same question, and why a couple of degrees of warming is an existential conversation for the industry rather than an academic one.
Altitude does most of the work
Within the belt, the single most predictive number is how high up the farm is. High means cold nights; cold nights mean the cherry ripens slowly; slow ripening means the seed has time to build sugars, acids and aromatic precursors, and grows dense doing it.
The course is blunt about the low end: a coffee grown at 600–800 m is boring. Not bad — boring. Push past 1,500 m and you start getting the fruit and spice notes that specialty coffee is sold on. This is why a good bag prints its altitude like a spec sheet, and why some origins can never win that argument: Brazil is mostly flat, and no amount of care changes what elevation it does not have.
What happens to the cherry
A coffee bean is a seed inside a fruit, and how you get the fruit off decides how much of it stays behind in the seed.
| Process | What happens | In the cup |
|---|---|---|
| Washed | Fruit pulped off, then fermented and washed clean before drying | Clean, bright, transparent. Nowhere for a mediocre lot to hide |
| Honey | Skin removed, sticky mucilage left on to dry | Between the two: round and sweet, still fairly clear |
| Natural | Whole cherry dried around the seed, fruit and all | Sweeter, jammy, heavier — and less intense, because the fruit blurs the origin's own edges |
And then someone roasts it
Roasting is the only one of the four you can watch happen, and it is the one that undoes the other three. As the roast develops, three things move at once: acidity falls from the very first crack, sweetness peaks somewhere around medium as sugars caramelise and then burns away, and bitterness climbs and never stops.
Underneath all three, origin character fades. Not linearly — it holds on through light and medium and then falls off a ledge. Past medium-dark, the flavours you are tasting were created in the roaster, not on the farm.
This is the uncomfortable part. At a dark roast, an exceptional high-grown Ethiopian natural and an ordinary low-grown commodity bean converge on nearly the same cup — smoke, body, bitterness. That is precisely why large roasters roast dark: it makes cheap coffee taste consistent, and it makes expensive coffee taste like nothing in particular. If you are paying for the farm, you need a roast light enough to still hear it.
Flavour builder
Pick a plant, a place, an altitude and a process, then roast it. Watch the last dial quietly take the other three away from you.
Pick an origin on the map or from the chips. The curve is redrawn for the coffee you have built.
Six things to try
- Erase a farm. Start where the builder opens — Ethiopia, natural, 1,900 m, light. Note the bars. Now drag the roast all the way to Italian, and then click through every origin chip in turn. Almost nothing moves. That is the whole chapter in one gesture.
- Find the ledge. Drag the roast slowly from light to dark and watch Farm voice rather than the bars. It does not fade evenly — it holds up through medium and then collapses between medium-dark and dark. Roasters live on that edge on purpose.
- Prove Brazil cannot be Kenya. Load Brazil and push elevation to its 1,200 m ceiling — the slider stops, because the country does. Now load Kenya and pull elevation down to its 1,400 m floor. Kenya at its dullest is still brighter than Brazil at its best. The end of the slider is the lesson.
- Change only the process. Hold Ethiopia at 1,900 m and click Washed → Honey → Natural. Fruit and sweetness climb, acidity and clarity give way. Same farm, same year, three genuinely different coffees.
- Meet robusta honestly. Switch species. The elevation slider collapses to the lowlands, because that is where the plant lives — then read the descriptor chips and the caffeine figure. Now put it at a dark roast and notice it is barely penalised for it.
- Hand one over. Build a coffee you would actually want, then follow the brew this link at the bottom of the panel — it opens the extraction simulator on a method that suits what you just built.
The recap
Four dials, one axis. Everything in the left column pushes a cup towards acidic and fruity; everything in the right pushes it towards earthy and bitter.
| Dial | Towards acidic & fruity | Towards earthy & bitter |
|---|---|---|
| Species | Arabica | Robusta |
| Elevation | Higher — 1,500 m and up | Lower — 600–900 m |
| Roast | Lighter | Darker |
| Origin | Africa, Central America | Indonesia, India |
Process sits slightly outside that grid, because it moves two things at once: washed pushes towards clarity and acidity, natural pushes towards sweetness and fruit. Neither is the bitter end.
Everything above is usually printed on the back, in this order: country and region, farm or washing station, variety, altitude in metres, process, and a roast date. If a bag gives you a roast date but no altitude and no process, that tells you something too — and if it only says "Italian roast", you already know how much of the farm you are being sold.
Unlike extraction, this one has no ground truth to check against. Each origin carries a base profile; elevation, process and species shift it; and roast interpolates the whole thing towards what a roaster tastes like, on a curve steep enough to swallow it entirely past medium-dark. Farm voice is measured rather than asserted — it is the distance between your cup and a featureless bean put through exactly the same roast.
But two naturals from neighbouring Ethiopian farms can taste nothing alike, and a great producer at 1,200 m will beat a careless one at 2,000 m every time. Treat this as "what these choices tend to do", never as "what this coffee tastes like".
Which leaves you with the useful version of the question. Not "is this a good coffee", but: which of the four decisions am I actually buying, and has the roaster left me enough of them to taste?