The olfactive pyramid, families and sub-categories

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.

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 and Iso E Super 180, so by hours alone both would file as heart to base materials, and both filings would be wrong: Hedione does a heart note's job and Iso E Super a base's. 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, the level the Scent Equalizer works at when you ask for more mossy and less smoky.

Across the catalog's 1,581 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. Open the bench, free, no account needed.