Beginner perfumer · 66
Five people diagnosed the same formula and none of them named the two biggest ingredients
· 20 min read
Someone posted a 14-material formula asking why their perfume was not strong enough. All five replies said versions of the same thing: too much orris, too much Javanol and Amber Xtreme, not enough top notes, unbalanced, it will turn into mud. I ran the arithmetic. After backing out the solvent carried in by two 10% dilutions, the actual aroma material is 3.161 g and the finished concentration is 19.38%. Sorted by weight, the largest item is Iso E Super at 31.6% of the aroma material, followed by Grojsman Plus at 19.8% — neither was mentioned by anyone. The orris named as "100% the reason" ranks fifth at 8.6%. And two of the top three read `medium` in the strength field.
About 6 minutes.
An hour ago someone posted this formula and asked why their perfume was not strong enough:
Iso E Super 0.999 g / Hedione 0.582 g / Javanol 0.300 g / Amber Xtreme 0.179 g / Grojsman Plus 0.626 g / Orris Imperial 0.273 g / Dihydromyrcenol 10% 0.096 g / Elemi 0.064 g / Green Mandarin 0.061 g / Fir 0.023 g / C-12 MNA 10% 0.022 g / Rose 0.019 g / Jasmine 0.015 g / Bergamot 0.008 g / perfumer's alcohol 13.042 g
Five replies, and the diagnosis was consistent:
"You have a huge overdose of Orris and a really big amount of both Javanol and AmberXtreme, which I suspect is crushing the blend. You also dont have a lot of top-notes to give you the diffusion you want. Overall it looks very unbalanced which usually turns into 'mud'." (frioke)
"Orris imperiale is 100% the reason. I had the same issue with orris givco. Make a batch with orris down to about 10% of what you have now and compare the difference." (Novel_Armadillo_6665)
Nobody did the arithmetic.
Back out the dilutions first
Two lines are marked 10%: dihydromyrcenol 0.096 g and C-12 MNA 0.022 g. The actual neat material is 0.010 g and 0.002 g; the remaining 0.106 g is solvent carried in by the dilutions.
| Weight | |
|---|---|
| Aroma lines as written | 3.267 g |
| Actual aroma material | 3.161 g |
| Solvent from dilutions | 0.106 g |
| Alcohol | 13.042 g |
| Total | 16.309 g |
| Finished concentration | 19.38% (20.03% on paper) |
19.38% is eau de parfum territory. It is not weak. So "not strong enough" is not a dilution problem.
(The denominator question is How many percent is three drops in 10 mL. The gap here is only 0.65 points because only two lines are dilutions.)
Sorted by weight
| Material | Actual g | % of aroma | % of bottle |
|---|---|---|---|
| Iso E Super | 0.999 | 31.6% | 6.13% |
| Grojsman Plus | 0.626 | 19.8% | 3.84% |
| Hedione | 0.582 | 18.4% | 3.57% |
| Javanol | 0.300 | 9.5% | 1.84% |
| Orris Imperial | 0.273 | 8.6% | 1.67% |
| Amber Xtreme | 0.179 | 5.7% | 1.10% |
| Elemi | 0.064 | 2.0% | 0.39% |
| Green mandarin | 0.061 | 1.9% | 0.37% |
| Fir | 0.023 | 0.7% | 0.14% |
| Rose | 0.019 | 0.6% | 0.12% |
| Jasmine | 0.015 | 0.5% | 0.09% |
| Dihydromyrcenol | 0.010 | 0.3% | 0.06% |
| Bergamot | 0.008 | 0.3% | 0.05% |
| C-12 MNA | 0.002 | 0.1% | 0.01% |
The top three add up to 69.8%.
And those three — Iso E Super, Grojsman Plus, Hedione — were not mentioned once in any of the five replies.
Orris Imperial, named as "100% the reason", is fifth at 8.6%.
Two of the top three read medium
Matched to the database by CAS:
| Material | Database record | Strength | Substantivity | Vapour pressure |
|---|---|---|---|---|
| Iso E Super | patchouli ethanone | medium | 172 h | 0.001 |
| Hedione | methyl dihydrojasmonate | medium | 72 h | 0.001 |
| Javanol | sandal cyclopropane | high (smell at 10% or less) | 400 h | 0.001 |
| Amber Xtreme | musk indenofuran | medium | 168 h | — |
| Dihydromyrcenol | dihydromyrcenol | medium | 16 h | 0.166 |
| C-12 MNA | 2-methyl undecanal | high (evaluate at 1.0% or less) | 388 h | 1.428 |
(Grojsman Plus and Orris Imperial are compounded bases and are not in the database. I counted why trade names go missing in Four kinds of nothing found.)
The two largest single materials in the formula both read medium, and together they are 50.0% of the aroma material.
Iso E Super and Hedione are both known for being transparent and for not taking over. That is not a flaw — it is exactly why they get used. But when half the aroma weight sits in two materials whose selling point is that they do not overpower anything, the result is not going to be force.
The only high in the top six is Javanol at 9.5%. frioke called the Javanol amount "really big" — on strength it is indeed the strongest of those six, but on weight it ranks fourth.
Putting a number on the top notes
frioke said "you dont have a lot of top-notes". That can be counted.
| Total | % of aroma | |
|---|---|---|
| The five woody/amber/base materials | 2.377 g | 75.2% |
| The five top-end materials (green mandarin, bergamot, fir, elemi, dihydromyrcenol) | 0.166 g | 5.2% |
The ratio is 14.4 : 1.
The vapour pressure column says it more bluntly. The two most volatile materials in the formula are:
- C-12 MNA: 1.428 mmHg
- Dihydromyrcenol: 0.166 mmHg
Iso E Super, Hedione and Javanol are all 0.001 mmHg.
C-12 MNA is 1,428 times more volatile than Iso E Super.
And those two are the only two he made up as 10% dilutions.
In real weight: C-12 MNA is 0.002 g — 0.1% of the aroma material, 0.01% of the bottle. Its strength field reads "high, recommend smelling in a 1.0% solution or less".
The one material in the formula that is both powerful and volatile is present at one part in ten thousand.
(Whether something at 0.1% can be smelled depends on its threshold, not its dose — that is Why 0.1% covers 30%. But the direction here is clear: he pushed the fastest-moving material down to a trace.)
Which kind of "not strong enough"
One person asked the right question:
"What do you mean by strong? Projection? Or longevity?" (whitenette)
The poster did not answer. And the two have opposite fixes.
- Not enough projection: the top and middle are thin. This formula fits that description (5.2% against 75.2%).
- Not enough longevity: the base is thin. This formula does not fit — Javanol 400 hours, Iso E Super 172, Amber Xtreme 168. The base is heavy.
So if he wants projection the diagnosis is right; if he wants longevity, the formula was never short of it.
(Strong on a strip and gone two steps away are different things: Strong on paper, gone at arm's length.)
About the arithmetic
Two people in the thread took aim at one line the poster wrote:
"Note: my formula is not AI generated, but it's a good tool to keep records of materials. It saves them for you and does all the maths 😎"
The reply:
"LLMs are bad at math. Which is why you see AI formulas people post that don't even add up to 100%. We have created a computer program that is bad at the one thing computers are good at." (berael)
The direction of that criticism is right, and it happens to describe what this article is doing. Every number above came out of a script, and every step of it can be re-run.
Formula arithmetic is the most verifiable part of this work — sums, dilution conversions, sorting. The most verifiable part is also the part you should least hand to something you cannot verify.
(Someone in the thread cited a 2023 study about ChatGPT's accuracy on prime-number checking falling sharply. I have not read that paper, so I am not repeating its numbers here.)
What to actually do
- Sort by real weight before diagnosing. The eye is pulled toward names that sound powerful, not toward mass.
- Convert 10% dilutions back to neat before sorting. The smallest line here came back as 0.002 g.
- Look up the strength field. Two
mediummaterials holding half the weight is not an accident, it is a choice. - Compute the top-to-base weight ratio. 14.4 : 1 is a number you can discuss; "unbalanced" is not.
- Settle whether "not strong" means projection or longevity first. The diagnoses point in opposite directions.
- Do not reason about two materials 1,428× apart in volatility using the same order of magnitude of dose.
- Do the arithmetic yourself. It is the only part of this work with a right answer.
What this doesn't establish
- I have not smelled this formula and have not made it. The whole article is weight arithmetic plus database fields.
- Grojsman Plus and Orris Imperial are compounded bases and are not in the database. Together they are 28.4% of the aroma material, so I know nothing about nearly a third of this formula. That is the biggest limit here.
- The split into "top five" and "base five" is mine, made from odour type and vapour pressure, not a database classification. Where elemi and fir belong is arguable.
- The strength field is the database's judgement, not my measurement, and it does not convert into dose.
- "Medium strength × 50% of weight = not strong" is an inference, not a calculation. The strength field has no unit you can multiply.
- Some vapour pressures are marked estimated, and the Amber Xtreme cell is empty.
- Whether C-12 MNA does anything at 0.01% depends on its threshold, which I did not look up. A low dose is not the same as no effect.
- The poster never defined "not strong enough", so both diagnoses can only be listed side by side.
- I did not verify the 2023 prime-number study, so none of its numbers appear here.