document
What a Formula Actually Looks Like
It is not a pyramid of notes. It is a table of materials and numbers, weighed to a thousandth of a gram.
Every fragrance you have ever bought was described to you as a pyramid. Top notes of bergamot and pink pepper, a heart of jasmine and rose, a base of sandalwood and musk. That description was written by a copywriter, usually after the perfume was finished, sometimes by someone who had not smelled it. It is not the document the perfume was made from.
The real document is dull to look at and much more interesting to read. It is a table. Left column, the name of a material. Right column, a number. At the bottom, the numbers add up to a round figure, almost always one thousand. There is no poetry anywhere on it. Everything a perfume is capable of doing to you is encoded in that table, and almost nobody outside the trade has ever seen one.
So here is one, in full, with the arithmetic left in.
Parts per thousand.
Perfumers work in parts per thousand. Not millilitres, not drops: parts by weight, out of a total of one thousand. It is a convention, and a good one, because it makes every number a tenth of a percent and makes two formulas instantly comparable regardless of batch size. Weigh it in grams and you have made a kilogram of concentrate. Weigh it in hundredths of a gram and you have made ten grams for a trial. The proportions are identical either way.
Below is a complete teaching formula, written in the ordinary house format. It is a small classical chypre: citrus at the opening, a floral heart carried mostly by dilutions, and a base built on the two Indonesian roots this journal has spent the last week on. Tap any row.
| Material | Dilution | Parts | % of conc. | Neat |
|---|---|---|---|---|
| Top | ||||
| Bergamot, Calabria | neat | 120 | 12.00 | 120 |
| Lemon | neat | 40 | 4.00 | 40 |
| Petitgrain bigarade | neat | 25 | 2.50 | 25 |
| Heart | ||||
| Geranium, Egypt | neat | 35 | 3.50 | 35 |
| Rose absolute | 10% DPG | 30 | 3.00 | 3.0 |
| Jasmine sambac abs. | 10% DPG | 20 | 2.00 | 2.0 |
| Clove bud | 1% DPG | 8 | 0.80 | 0.08 |
| Base | ||||
| Patchouli, Sulawesi, aged | neat | 150 | 15.00 | 150 |
| Iso E Super | neat | 120 | 12.00 | 120 |
| Vetiver, Java | neat | 90 | 9.00 | 90 |
| Musk, Habanolide | neat | 80 | 8.00 | 80 |
| Ambroxan | 10% DPG | 60 | 6.00 | 6.0 |
| Oakmoss absolute | 10% DPG | 45 | 4.50 | 4.5 |
| Labdanum resinoid | 10% DPG | 40 | 4.00 | 4.0 |
| Benzoin resinoid, Sumatra | 10% DPG | 22 | 2.20 | 2.2 |
| Vanillin | 10% DPG | 15 | 1.50 | 1.5 |
| Solvent | ||||
| Dipropylene glycol (DPG) | neat | 100 | 10.00 | 100 |
| Total | 1000 | 100.00 | , | |
The pyramid is written after the perfume. The table is written before it. Only one of them ever touched the bench.
You cannot weigh eight thousandths of a percent.
Look back at the clove bud line: 0.08 parts neat in a thousand. If you were compounding ten grams for a trial, that is 0.0008 grams. No bench scale will do it, and even if one would, a single extra crumb would wreck the trial. So perfumers do not add powerful materials directly. They add them as solutions, pre-made at ten percent, one percent, sometimes a tenth of one percent, in an odourless solvent, and they keep those solutions in front of them permanently.
This is also how you learn a material. Almost nothing is comprehensible neat. Smell most aroma chemicals undiluted and you get a headache and a false impression. The dilution ladder is not just a weighing convenience, it is how the material is heard at the right volume.
Nobody gets it right on the first sheet.
A formula is not written, it is iterated. Every version is weighed, left to sit, smelled on paper and on skin at intervals, annotated, and then rewritten with a new number in the corner. Perfumers call these mods, short for modifications, and they are numbered without ceremony. Mod 1. Mod 2. Mod 17. It is entirely normal for a finished commercial fragrance to sit somewhere between the fortieth and the four hundredth attempt, and the difference between mod 38 and mod 41 is often a single number moved by two parts.
Below is a plausible, compressed history of the formula above. Step through it.
Two things in that sequence are worth naming. First, most of the moves are subtractions. The instinct of every beginner is to add a material when something is missing; the instinct of every experienced perfumer is to remove one. Second, the trial that finally works is usually not the most complicated. Mod 43 has fewer materials than mod 19.
The concentrate is not the perfume.
Everything on that sheet adds up to a thousand parts of concentrate, also called the compound or, in the trade, simply the juice. It is not wearable. It is a thick, intense, faintly syrupy liquid that would be unpleasant on skin and is not what goes in the bottle.
The concentrate is then dissolved in ethanol, and the proportion decides what the label says. Roughly eight to twelve percent gives an eau de toilette. Fifteen to twenty percent gives an eau de parfum. Twenty-five to thirty and you are in extrait territory. That is the whole secret of those categories: the same formula at different strengths, with small adjustments, sold at different prices.
Then it waits. The blend is left to macerate, commonly for several weeks, in the dark and cool, while the materials react and settle and the raw edges of the mixture round off. Then it is chilled and filtered to remove the waxes and haze that natural materials throw out, then rested again. A fragrance shipped straight after compounding smells noticeably harsher than the same fragrance shipped six weeks later. This is also why the sample you smell in the lab and the bottle on the shelf are, strictly, not the same liquid.
About one dollar.
Now put prices against the sheet. Using ordinary trade figures for materials of this quality, the formula above costs somewhere in the region of one hundred and twenty-five dollars per kilogram of concentrate. A fifty millilitre eau de parfum contains roughly eight grams of that concentrate.
Roughly one dollar. That is the honest arithmetic, and it is the number people find upsetting, so it is worth being precise about what it does and does not mean. It does not mean a perfume should cost a dollar. The bottle, the cap, the box, the filling, the shipping, the duty, the retailer's margin, the returns, the sampling, the photography and the years of unpaid work behind mod 43 are all real costs, and in most of the industry the advertising costs more than everything else combined. What the number means is simply this: almost none of what you pay for a fragrance is the fragrance.
Which is exactly why the composition of that one dollar matters so much, and why a house should be able to tell you what is in it.
Then a spreadsheet tells you what you may keep.
Before a formula can be sold anywhere serious it goes through compliance, and compliance is not a formality. The industry body, IFRA, publishes standards that prohibit some materials outright and cap many others at a maximum percentage in the finished product, with different limits depending on whether the thing is a perfume, a body lotion or a candle. Separately, a list of fragrance allergens must be declared on the packaging above certain thresholds, which is why the back of a box carries a paragraph of chemical names most people assume are additives. They are not additives. They are usually the natural components of the flowers.
The consequences are real and often sad. Oakmoss, the material that makes a chypre a chypre, is capped so tightly that the entire genre had to be rebuilt. Countless people who say a favourite fragrance was ruined are correct, and the cause is usually not corporate cheapness but a regulatory limit arriving mid-life. It also explains a strange fact of modern perfumery: a formula written in 1970 would today be illegal to sell, and its reformulation is a different perfume wearing the same name.
Two materials at a time, for years.
There is a training method behind most of this, and it is nearly a century old. Jean Carles, the perfumer behind Dior's Miss Dior and a founder of the school at Roure in Grasse, built a systematic curriculum from a simple observation: you cannot learn six hundred materials by smelling six hundred materials. You learn them in contrast.
Students work through classified tables, first comparing whole families against each other, then materials within a family, then building two-material accords in a ladder of proportions, 90:10, 80:20, 70:30 and so on, to find the point where the pair stops being two smells and becomes one. Only then are third and fourth materials introduced. Carles reportedly continued teaching after losing his own sense of smell, working entirely from the structural logic he had built. The tables are still revised whenever regulation changes a permitted dose.
It sounds mechanical and it produces the opposite. What the method actually trains is the ability to hear a mixture as a shape rather than as a list, which is the difference between someone who can name notes and someone who can write that table at the top of this page.
Formulas are property, and property is secret.
You will not find the real formula for any famous fragrance, because the formula is the asset. Unlike a recipe, it is generally not patented, since a patent would require publishing it. It is kept as a trade secret, often known to a handful of people, sometimes with materials disguised in the house's own internal codes so that even the compounder does not know what is being weighed.
That secrecy is leakier than it used to be. Gas chromatography coupled to mass spectrometry will separate a fragrance into its components and identify most of them, which is how the dupe industry functions and why a competent lab can produce a recognisable copy of a bestseller within weeks. What GC cannot reliably read is the quality tier of a natural, the age of a patchouli, or the reason a number is 47 rather than 50. The copy is usually correct and slightly dead.
Which leads to the small confession this house owes you. We publish a teaching formula, not ours. Ours change: we make in small batches, we accept variation between them, and we would rather a material be interesting than identical. A formula in this house is a set of intentions with numbers attached, not a specification to be hit forever.
Sixteen lines, one thousand parts, about a dollar of oil. Everything else you are paying for is the story around it, and the story should at least be true.
The other formula, the one nobody wrote down.
Everything above is a Western technical artefact: a document, in numbers, owned by somebody, protected as property. It is not the only way a fragrance formula can exist.
This country has had one of its own for centuries. Seven flowers, in fixed proportions, assembled for birth and marriage and death, sold by the bag at the market gate and made fresh every single time. Nobody holds the intellectual property. Nobody has ever written it as a table. It is a formula transmitted entirely by hand and by memory, which is a description of the thing this house is named after.
Kembang Setaman
Seven flowers, one bowl of water. The oldest working perfume formula in the archipelago, composed for birth, marriage and death, and never once bottled.
The formula above is an original teaching example written for this essay. It is a genuine, workable structure and it is not any commercial fragrance. Prices are representative trade figures and move constantly.
- Concentration bands for eau de toilette, eau de parfum and extrait, and the use of DPG: A guide to standard formulations.
- Working dilutions, weighing by weight rather than volume, and the C1V1 = C2V2 calculation: How to dilute perfume materials.
- The Jean Carles method, the classified tables and accord ladders: A Method of Creation and Perfumery, Jean Carles, 1961; The Jean Carles method, olfactory training charts.
- Formulation practice and the perfumer's bench: The Perfumer's Wizard guide.
- Related reading in this journal: Patchouli and Akar Wangi, the two most expensive lines on the sheet.