A bag of specialty coffee promising jasmine, stone fruit and brown sugar invites reasonable scepticism. It reads like marketing, and quite often it is. But underneath those descriptors sits a genuine body of sensory science, and the difference between a tasting note grounded in that work and one invented by a copywriter is real. This article covers where the vocabulary came from and what it can and cannot do.
The problem the lexicon solved
Until the mid-2010s, coffee's flavour vocabulary was largely traditional. The Specialty Coffee Association's original flavour wheel dated from 1995 and was not built from systematic data. Terms were inherited, locally variable, and undefined. If a buyer in Seattle and a producer in Ethiopia both described a lot as winey, there was no way to establish they meant the same thing, and no reference either could point at.
That is a commercial problem before it is an academic one. Coffee is bought and sold on quality descriptions across long distances, and a vocabulary without shared definitions makes contracts and feedback unreliable. A producer told their lot tastes earthy cannot act on that without knowing what the word denotes.
The lexicon
The response was the World Coffee Research Sensory Lexicon[1], developed with the Sensory Analysis Center at Kansas State University, a research group specialising in descriptive sensory analysis across food categories. The lexicon identifies 110 distinct flavour, aroma and texture attributes present in coffee.
Its defining feature is that each attribute comes with a physical reference and an intensity scale. The term for a given flavour is anchored to an actual substance a panel can smell and taste, with a defined intensity value. Trained panellists calibrate against these references, which means that when two panels in different countries record the same attribute at the same intensity, they are describing something genuinely comparable. This is standard practice in descriptive sensory analysis and it is what separates the method from casual tasting.
It is worth being clear that these panellists are trained instruments, not enthusiasts giving opinions. Descriptive analysis deliberately excludes preference. A panel records what is present and how strongly, not whether it is good.
Turning a list into a wheel
A list of 110 attributes is scientifically useful and practically unusable at a counter. The step from lexicon to wheel is where the second piece of research comes in.
What descriptive analysis actually involves
It is worth describing how a trained sensory panel works, because it is very different from what people imagine when they hear coffee tasting, and the difference is what gives the method its authority.
Panellists are screened for sensory acuity and then trained, often over dozens of hours, against physical reference standards. The training's purpose is calibration rather than education: the aim is that when two panellists record an attribute at intensity 6, they mean the same thing. Panels are periodically re-calibrated against the references because sensory judgement drifts.
Crucially, descriptive analysis deliberately excludes preference. Panellists record what is present and how intensely, not whether they like it. That separation is what makes the output usable as data. A panel can report that a coffee is intensely earthy without any implication about whether that is desirable, and the buyer decides what to do with the information.
This is why a lexicon-derived tasting note and a marketing tasting note are different objects even when they use the same words. One is the output of a calibrated instrument; the other is a sentence someone wrote to sell a bag.
Spencer, Sage, Velez and Guinard published the method in the Journal of Food Science in 2016[2]. Seventy-two experts took part in a modified online free sorting task, grouping the lexicon's flavour attributes by similarity according to their own judgement. Free sorting is a well-established sensory technique that captures how people naturally organise a domain without imposing categories in advance.
The resulting data were analysed with agglomerative hierarchical clustering to determine the clusters and the levels within them, and with multidimensional scaling to determine how those clusters should be positioned relative to one another around the wheel. That is why related flavours sit adjacent on the finished wheel and unrelated ones sit opposite: the geometry is a statistical output, not a design decision.
The 2016 Coffee Taster's Flavor Wheel that emerged from this is now the standard reference across the specialty industry, and the collaboration behind it, between World Coffee Research and the Specialty Coffee Association, has been described as the largest piece of coffee flavour research undertaken.
What it is good for, and what it is not
Three honest limits. First, the wheel describes what is present, not what is good. A well-defined attribute can be one nobody wants; the lexicon includes plenty of negative attributes and treats them with the same rigour as the pleasant ones.
Where the flavours come from
It is worth connecting the sensory vocabulary to its causes, because the descriptors are not arbitrary and several of them trace back to decisions made long before roasting.
Species and variety set the outer bounds. Salojärvi and colleagues' work on the Arabica genome[3] showed how narrow Arabica's genetic base is, and genetics determine which precursor compounds a bean can contain at all. Processing method, whether the fruit is stripped early or dried on the seed, substantially shapes fruitiness and acidity. Growing altitude affects the density and sugar development of the bean. Roasting then converts those precursors, through the Maillard reaction and caramelisation, into most of the aromatic compounds you actually smell.
So a tasting note is downstream of a chain of decisions running from the genetics of the plant to the moment the roast was stopped. That is worth knowing because it explains why the same origin can taste completely different from two roasters, and why blaming or crediting any single step usually misattributes the cause. Our guide to processing methods and roast levels guide cover two of the largest links in that chain.
Second, the descriptors are analogies rather than ingredients. A coffee described as blueberry contains no blueberry. Certain volatile aromatic compounds in that coffee overlap with compounds in blueberries, and a trained panel recognises the resemblance. This is the point most sceptics get stuck on, and the answer is that the note is a statement about aroma chemistry, not about contents.
The chemistry behind that is not hand-waving. Coffee contains hundreds of volatile aromatic compounds, and many are shared with other foods: compounds in the same families as those found in berries, in cocoa, in caramel and in florals genuinely appear in roasted coffee, produced during roasting from precursors in the green bean. When a panel records blueberry, they are recognising an aromatic profile that overlaps with blueberries, which is a chemical fact rather than a poetic gesture.
Third, a note on a retail bag is not necessarily lexicon-derived. Nothing prevents a roaster writing whatever they like. The tasting notes attached to formally scored lots and competition coffees are far more likely to trace back to trained evaluation than the ones on a supermarket shelf.
How this connects to grading
The sensory framework underpins the specialty grading system. Coffees are evaluated by trained cuppers against a standard protocol scoring attributes including fragrance and aroma, flavour, aftertaste, acidity, body, balance, uniformity, clean cup and sweetness, producing a score out of 100. The commonly used threshold for specialty grade is 80 points.
The economics riding on a sensory score
It is worth being explicit about how much money moves on these judgements, because it explains why the research was funded and why standardisation mattered enough to build a lexicon for.
Commodity coffee is priced off a global futures market that treats it as an undifferentiated bulk good. Specialty coffee is priced separately, and the premium is substantially driven by cup score. A lot scoring 84 and a lot scoring 79 from the same region in the same year can command very different prices, and the difference is determined by trained people tasting them.
That makes the sensory framework a piece of economic infrastructure rather than an aesthetic exercise. It also means access to it is a form of market access. A producer whose coffee has never been formally evaluated cannot sell into the specialty market at specialty prices regardless of how good the coffee is, because the price mechanism runs through the score.
This is directly relevant in the Philippines, where, as covered in our article on the Philippine coffee industry, most production is Robusta destined for instant manufacture and farmgate prices sit well below what specialty Arabica commands. Cupping infrastructure, trained Q graders and access to evaluation are part of what stands between a smallholder and a better price.
That number determines what a lot is worth. Green coffee pricing above the commodity baseline is substantially driven by cup score, which means the sensory vocabulary has direct financial consequences for producers. This is not an abstraction for Philippine farmers: as covered in our article on the Philippine coffee industry, most domestic production is Robusta destined for instant manufacture, and access to specialty pricing depends on the ability to have coffee formally evaluated at all.
Using it as a drinker
The wheel's most useful property for an ordinary drinker is its structure. It runs from broad categories at the centre to specific ones at the rim, and you are meant to work outward as far as you honestly can and then stop. Getting to fruity is a real observation. Pushing on to fruity, then berry, then blueberry is a skill that develops with practice, and stopping at fruity is not a failure.
It is also worth giving yourself permission to disagree with a bag. Descriptors are produced by a specific person or panel tasting a specific batch, brewed a specific way. Your water, your grinder, your brew ratio and your palate all differ. If you cannot find the jasmine, the most likely explanations are that your brew is extracting differently, that the coffee has aged since it was described, or simply that you are not yet trained to isolate that note. None of those mean the note was invented.
The practical exercise that builds this fastest is comparison. Tasting two coffees side by side makes differences obvious that are invisible in a single cup, because the palate is far better at relative judgements than absolute ones. Two contrasting origins, or a light and a dark roast of the same bean, will teach more in ten minutes than a month of drinking one coffee attentively. Our roast levels guide and taste-based drink guide are useful companions, and a Marikina roaster like Franken Coffee Bar + Roastery is a reasonable place to try origins side by side.
The summary
Coffee's flavour vocabulary rests on the World Coffee Research Sensory Lexicon, 110 attributes each anchored to a physical reference and an intensity scale, developed with a university sensory research centre. The Coffee Taster's Flavor Wheel derived from it was built by having 72 experts sort those attributes and then deriving the groupings and geometry statistically. The system describes rather than evaluates, its descriptors are aroma analogies rather than ingredients, and its use on any given retail bag is entirely voluntary. Where it is applied properly, it turns what looks like florid marketing into something two people on opposite sides of the world can actually agree on.
That is a genuine achievement for an industry that spent most of its history describing coffee as strong or weak.
Used well, it is the difference between knowing you liked a coffee and knowing why.
The short version for a drinker: the wheel is a real research output, the words are meant to be analogies rather than ingredients, and the vocabulary is worth learning mainly because it makes you pay attention. Naming what you taste is most of what improves your palate.
Everything else follows from paying attention with a shared vocabulary.
The wheel is a prompt for that attention more than a test to pass.
Nobody is grading you on finding the jasmine.