Perfume drydown: what it is and how to read it

What is a perfume drydown?

A perfume is not one smell but a schedule of smells. Put a formula on a blotter and read it at ten minutes, at two hours and at hour eight, and you meet three different compositions: the bright opening gone, the middle blooming, the quiet residue that is still there the next morning. The drydown is that schedule, the way a composition changes as its materials evaporate at different speeds.

The classic way to teach it is the pyramid: top notes carry the opening minutes, heart notes the first hours, base notes the rest of the day. The pyramid is a good filing system, and it is not physics. What actually happens is evaporation. Every material leaves the skin or the strip at its own rate, and a perfume's shape over time is all of those departures added together.

The filing and the physics part company exactly where a material is tenacious without being deep. Hedione reads 72 hours on a strip and Iso E Super 180, so the hours would file both as heart-base, and both filings would be wrong: the first does a heart note's job and the second a base's. A blotter measures how long a molecule goes on evaporating; a perfumer files by the job the material does. An honest model keeps the two separate, which is what PerfuMate does: the hours drive the curves, the top, heart or base band stays the perfumer's filing, and the label can move without the physics underneath it moving.

Why does a perfume change when the grams do not?

Here is the idea the whole reading hangs on. Weigh a formula and the grams never move again: the bergamot is the same three grams in the bottle at hour eight that it was at minute one. What changes over the life of a perfume is not how much of a material exists but how much of it is reaching a nose, a function of how fast it leaves, how strong it is, and how long it has already been leaving. Reading a formula by weight tells you what you own. Reading it by impact tells you what it smells like, and only the second reading has a clock worth turning.

So PerfuMate keeps the two views deliberately apart. The amount view is a static composition of what is in the bottle, with no clock on it at all; the drydown slider belongs to impact mode alone. Underneath, the calculation is honest about being a calculation: handed a clock in amount mode it will decay the physical mass rather than refuse, which is what the Scent Equalizer's three moments and the amount-mode PDF filmstrip use. The screens make the choice; the maths does not pretend the other reading is impossible.

What does the model read from each material?

PerfuMate models a formula from the behaviour of its materials rather than from the formula as a whole. For every line on the bench the model reads five things:

1. Dilution-adjusted aromatic mass. Every dose is corrected for the strength of the bottle it was actually weighed from, so a line poured from a 10% solution counts for a tenth of what the same gram of neat material would. This is the weight everything else multiplies.
2. Family and sub-category dominance order. The catalog does not store a hand-typed percentage per material; it stores the order its olfactive families come in, and one shared curve turns that order into shares: the primary family takes 6 parts, the second 3, the third 2, the rest 1 each, with a family's share split across its sub-categories. The same curve feeds the profile chart, the pyramid and the colour of the material's swatch, so they agree by construction rather than by coincidence.
3. Pyramid tier. The top, heart or base band the material is filed under. A few hundred materials state it outright in the catalog, the rest derive it from their hours, and the band picker in the material editor lets the perfumer overrule either.
4. Substantivity. How many hours the material is still readable on a blotter, and it drives exponential decay: 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.
5. Perceived impact. Evaporation flux times potency, compressed the way perception actually compresses, times a bloom for how much of the material is airborne yet. A potency rating sits on every one of the catalog's 1,581 materials without exception.

The drydown is those per-material curves summed over 24 hours, and the longevity estimate is where the summed curve falls under the perception floor.

How do I read the drydown chart?

The timeline sits under the scent profile as a slider from hour 0 to hour 24, with families nested alongside their sub-categories and every sub-category carrying its own weight. Drag it, and the picture at each hour is what the model says is reaching a nose at that hour.

The chart is absolute: notes genuinely fade rather than being renormalised to keep the picture full. This matters more than it sounds. A renormalised chart is always 100% of something, so a perfume that has collapsed to a whisper of musk looks as busy at hour twelve as it did at minute one, and the thing you opened the chart to learn, how much is actually left, is exactly what it hides. An absolute chart is allowed to go quiet, because perfumes do.

As far as we have looked, none of the tools on the comparison page ship an interactive drydown timeline at all. Their claims were re-read from their own pages on 2026-08-17, so that is "none of the tools we compared", not a claim about every product on earth.

How far can the drydown be trusted?

The hours are sourced, not measured by us. Hundreds of materials carry a sourced substantivity figure, the app labels every one of them an estimate on the face of the number, and where none exists it estimates from family and band. A blotter is not an exact science and the product says so wherever it prints one.

Every predicted number carries a computed confidence: the app weighs how much of a formula's aromatic mass rests on measured values and prints that beside the prediction, from 40% when nothing is measured up to 95% when everything is. It never prints 100, because this is still a model of a blend on skin, and it cannot see skin chemistry, weather, dose, or what the materials do to each other. After a wearing you can record what actually happened, and the app then shows its own error on your own formulas; nothing is adjusted automatically and nothing is uploaded. The full account is at how the predictions are made.

One more honest limit: there is no physical chemistry on a catalog row, no molecular weight, no vapour pressure, no boiling point. The volatility model is built from sourced blotter figures and family, and it does not pretend otherwise.

What do you do with a drydown chart?

You look for the hours where the picture thins, and for an opening that promises more than the rest of the wear keeps.

Spotting a hole at hour four. Drag the slider to hour four and look at what is still standing. If the opening families have burnt off and the base has not yet risen, the summed picture thins to almost nothing: the classic mid-wear hole, and it is invisible in an amount view, because the grams that would fill it are all still in the bottle. The fix is usually a heart material with a longer substantivity, and because the chart is absolute you watch the hole actually fill rather than the picture merely rearranging itself.

Spotting a top-heavy opening. At hour 0, notice how much of the chart's height belongs to families whose curves are tall early and gone by the second hour. Then drag to hour one and see what survives. If most of the opening does not, the perfume makes a promise the drydown will not keep, and the chart shows you that before a blotter has to.

And when the reading turns into a wish, "more mossy at hour eight", the Scent Equalizer takes exactly that sentence and returns concrete gram moves chosen from materials you already own, composed as one idea rather than patched line by line.

The drydown, the slider and the confidence figure are all free with a free account. The paid version is a one-time purchase, not a subscription, and what they sell is room rather than tools. The fastest way to understand any of this is to load a formula and turn the clock: Open the bench and read your own formula at hour four.