How to test a perfume formula

Test a perfume formula on paper strips at fixed time marks, compare versions as blind pairs, change one thing between trials, and write down what you smelled before you decide anything. Skin comes last, because skin is the noisiest instrument you own and it should confirm a result, not produce one.

Should I test on strips or on skin?

Strips first, skin last. A blotter strip has no temperature, no sweat and no yesterday's shower gel on it, so what you smell is the formula and nothing else. It also lets you run six versions side by side, which skin cannot do: you have two wrists and they share an air space.

Skin answers a different question. It tells you how the finished perfume wears on a warm surface for a whole day, and whether the base you approved on paper turns sour, flat or loud against a body. Take a formula to skin once it has already passed on strips, and treat skin as the final exam rather than the homework.

Dip the strip, do not soak it. Label it with the version and the time before you dip, because a strip with no label is a strip you will confuse with its neighbour by the second hour.

At what times should I smell a strip?

Smell at fixed marks and the same marks every time, so trials can be compared. A working set:

• 30 seconds: the opening. Solvent flash is still there; do not judge much yet.
• 10 minutes: the top notes as a wearer meets them.
• 1 hour: the heart. This is where most formulas are actually decided.
• 4 hours: the turn into the base.
• Next morning: what is left. If nothing is left, that is a finding too.

Between marks, put the strips somewhere with still air and walk away. Smelling a strip every two minutes tires your nose and tells you nothing the marks would not.

What is a blind pair and why does it matter?

A blind pair is two strips, the current version and the previous one, labelled on the underside by someone else, or by you with the strips shuffled after labelling. You smell both at each mark and write which you prefer and why before you turn them over.

The reason is simple: you already know which version you spent the afternoon on, and you will prefer it. Blind pairs are the only cheap way to find out whether a change helped or whether you merely wanted it to. If you cannot tell the two apart at any mark, the change did nothing and the version can be discarded without regret.

Why one change per trial?

Because a trial with two changes has no answer. If you raise the musk and lower the citrus in the same version and the result is better, you do not know which move did it, and the next version inherits a guess. One change per trial is slow on the day and fast over a month, because every version teaches one clean fact.

A change is one line moved by one step: a material added, removed, or dosed up or down. Doubling a dose is a bigger step than most formulas need; a move of a quarter or a half is usually enough to hear.

A worked example

Take a small woody sketch: Iso E Super 40 parts, Vetiveryl acetate 20, Bergamot Essential Oil 15, Hedione 20, Ambroxan 5. Call it version 1.

At the one hour mark the strip reads pleasant and faceless. The suspect is the Iso E Super, which is smooth by nature and at 40 parts is smoothing everything around it. Version 2 changes one line: Iso E Super down to 30, the total no longer summing to 100, which is fine on strips because ratios are what you are hearing. Blind pair at 10 minutes, 1 hour and 4 hours. If version 2 wins at the hour and loses at 4 hours, you have learned that the Iso E Super was also carrying the drydown, and version 3 is a base question, not a heart question. Perhaps Vetiveryl acetate up to 25. One line again.

How do I write a trial down?

Write at the strip, not from memory in the evening. The note needs five things and no more:

1. The version number and what changed from its parent, in one line.
2. The time mark.
3. What you smelled, in plain words: "sharp", "soapy", "gone", "the bergamot is still here".
4. Which of the blind pair you preferred at that mark.
5. What the next single change will be, decided before you read the labels.

A trial that is not written down did not happen. A week later the strips are dry and identical, and all you will have is the note.

PerfuMate keeps this loop in one place with your own shelf: every version records the version it came from, so the one-change rule is visible in the evolution tree, the drydown chart shows what each version should do at each time mark before the strip dries, and trial notes and the maceration journal live next to the formula they describe. The AI is handed the catalog and your shelf on every run and is never trained on them. See how the bench works.