The olfactive pyramid, families and sub-categories

By Omer Siton, founder of PerfuMate · Updated 2026-10-15

Every perfume description leans on the same three words: top, heart and base. They are useful, and they are also widely mistaken for a law of chemistry when they are really a filing system. This page explains what the pyramid is, where it stops being physics, what olfactive families and sub-categories add underneath it, and exactly how PerfuMate turns a material's character into the numbers on a chart.

What is the olfactive pyramid?

The olfactive pyramid sorts materials into three bands by the part of a perfume's life they carry: top notes lead the opening, heart notes hold the middle hours, base notes are what remains at the end. In PerfuMate every odorant in the catalog carries a top, heart or base band. Solvents and carriers do not: they have no smell to file, so they stand outside the pyramid and take no share of a formula's character.

The band is one of two time-related facts on a material's row. The other is substantivity, how many hours the material is still readable on a blotter: sourced where a source exists, estimated from family and band where none does, and labelled an estimate wherever it is printed. The two are deliberately kept separate, because they answer different questions.

What is a perfume note?

A note is one smell impression in a perfume, and the word is used three ways: for a material ("a bergamot note"), for a facet you can name ("a powdery note") and for the stage of the drydown it belongs to ("a top note"). The notes listed for a perfume are a description written for the wearer, not its ingredient list.

The difference matters when you make perfume. "Rose" among a perfume's notes can come from rose oil, from a rose accord of a few molecules, or from neither: PerfuMate's published Rose Accord builds its rose from citronellol, geraniol, phenethyl alcohol and seven more lines, with no rose oil at all. A formula is written in materials and grams; the notes are what it smells like.

What are top notes?

Top notes are the most volatile materials in a perfume: they make the first impression, in the first minutes, and are mostly gone within the first hour or so on skin. Citrus oils, light herbs and small, fresh molecules are typical.

The catalog shows how fast they go. On a smelling strip, Lemon Essential Oil reads about 4 hours, Bergamot Essential Oil about 8, Pink Pepper Essential Oil about 10 and Linalool about 12, and skin holds them for less. That is why a perfume needs something to bridge from the top into the heart, or it goes hollow after twenty minutes (bridging the gap).

What are heart (middle) notes?

Heart notes are the body of a perfume: the materials that emerge as the top fades and carry its character for the next several hours, usually florals, spices and soft aromatic materials.

Rose alcohols such as Geraniol, about 60 hours on a strip, and Citronellol, about 56, are classic heart materials. Hedione shows why the band is a perfumer's filing and not only a clock: it reads about 72 hours on a strip, long enough to pass for a base note, and it is filed heart because its job is to carry the middle of a composition with a transparent jasmine.

What are base notes?

Base notes are the slowest materials in a perfume: woods, musks, ambers, resins and vanilla that stay on skin for hours and on a strip for days or weeks. They are the drydown, what is left at the end of the day, and the main reason a perfume lasts.

On a strip, Ambroxan reads about 400 hours, Iso E Super about 280 and Ethylene Brassylate about 208. Base materials that also hold a little of the lighter materials back with them are what perfumers call fixatives (what a fixative is).

How long do top, middle and base notes last?

On a paper strip, the typical top note lasts about 7.5 hours, the typical heart note about 68 hours and the typical base note about 300 hours: the medians of the sourced blotter figures in PerfuMate's catalog, grouped by the band each material page prints. On skin, worn in a perfume, the top carries the first minutes to an hour or so, the heart the next few hours and the base the rest of the day.

Two cautions. The bands overlap: Olibanum Essential Oil is filed top for its fresh, peppery opening and still reads on a strip after 284 hours, which is why PerfuMate reads each material's own hours rather than its band. And skin is not paper: PerfuMate's model takes a material's life on skin as a fraction of its strip life, a larger fraction for the fleeting ones, so 4 hours on paper is about 1 hour on skin, 40 hours about 7.5 and 180 hours about 26. The strip figures are sourced, not measured by PerfuMate, and the app labels each one an estimate.

Which note in a perfume is the strongest?

None of the three bands is stronger by nature: strength belongs to each material, and across the catalog the median impact is 6 out of 10 in the top, the heart and the base alike (the catalog, counted). What differs is speed.

Top notes evaporate fastest, so they put the most material into the air in the first minutes and are the loudest thing at the opening: in PerfuMate's model, at equal mass, bergamot gives off about eight times the evaporation flux of galaxolide. Base notes are the last thing left, which is why they seem strong late in the day, and a strong base material dominates at a small dose: Ambroxan is rated 7 out of 10, Ethylene Brassylate 2 (what makes a perfume project).

Are top, heart and base real, or a filing system?

The hours are the physics. They drive the drydown, the wear model and the longevity estimate: a short substantivity gives a curve that is tall early and gone by the second hour, a long one stays low and is still there at hour twelve.

The band is the perfumer's, because a blotter cannot answer the question it is being asked. Hedione reads 72 hours on a strip, so by hours alone it would file as a heart to base material, and that filing would be wrong: Hedione does a heart note's job. Iso E Super, at 280 hours, is one where the hours and the job agree: base. A blotter measures how long a molecule goes on evaporating; a perfumer files by the job the material does in a composition, and the two part company exactly where a material is tenacious without being deep.

So the catalog states the band outright for a few hundred materials, derives it from the hours for the rest, and the band picker in the material editor lets you overrule either. The label moves; the physics underneath it does not.

What is the difference between a family and a sub-category?

The pyramid says when a material speaks. Families say what it says. A family is the broadest statement of a material's character, and a sub-category is the finer grain inside it.

Across the catalog's 2,211 materials there are 21 olfactive families and over two hundred family and sub-category pairs in live use, with thousands of sub-category assignments in all, close to five per material. Most materials sit in more than one family at once, and they carry those families in a definite order: which character leads, which is secondary. That is the point of the classification: a material is not a label, it is a mixture of characters in a definite order. The few that carry no olfactive family at all are solvents and diluents, filed in buckets that carry no smell and take no share of a formula's character.

You can see all of this on any material page: olfactive families with percentage dominance, beside the odour description, impact and substantivity. Hedione and Iso E Super are good places to start, and the whole catalog index is free to browse, grouped by family.

How does an order of families become a percentage?

The dominance percentages on those pages are not typed in by hand, and that is deliberate. The catalog stores the order a material's families come in, and one shared curve turns that order into shares: the primary family takes 6 parts, the second 3, the third 2, and the rest 1 each. A family's share is then split across its sub-categories, so each sub-category lands on the chart with its own weight rather than the family arriving as one block.

That single curve feeds four things at once: the scent profile chart, the pyramid, the colour of the material's swatch and the genre title. All four agree by construction rather than by coincidence, because they are one calculation instead of four maintained side by side.

Inside a whole formula, each material's share is then weighted by its dilution-adjusted aromatic mass, so a line weighed from a 10 percent solution counts for a tenth of what the same gram of neat material would.

What does that look like on a real material?

Take a material filed under three families, in order. The curve gives the first 6 parts, the second 3 and the third 2, which is 11 parts in all, so its character reads as roughly 55 percent the leading family, 27 percent the second and 18 percent the third. If the leading family carries two sub-categories, that 55 percent is split between them, and each carries its own weight on the chart.

Now weigh 2 grams of it into a formula from a 10 percent dilution. Its aromatic mass is 0.2 grams, and it is that figure, not the 2 grams in the bottle, that scales its 55, 27 and 18 against every other line. Strength in the bottle, not volume poured, decides how loudly each character speaks in the profile.

The drydown timeline then adds the clock. Each material's contribution decays exponentially against its own substantivity, scaled by a perceived impact: evaporation flux times potency, compressed the way perception actually compresses, with a bloom for how much is airborne yet. Families stay nested with their sub-categories along the whole 0 to 24 hour slider, and the chart is absolute: notes genuinely fade rather than being renormalised to keep the picture looking full.

How far to trust the result, and where the confidence figure beside every prediction comes from, is answered plainly at how the predictions are made.

The quickest way to see the curve at work is on a formula of your own: add a few materials, open the scent profile, and drag the slider. Start a formula, free. No account needed.