Formula cost and sourcing: the expensive material is rarely the problem — not being able to buy it is

Faced with "orange blossom absolute, 10 mL, £120", a beginner usually does one of two things: gives up on the material, or buys it and then cannot bring themselves to use it.

Both reactions rest on the same mistake: treating unit price as cost.

Cost is dose times price

There is one formula for formula cost: each material's percentage times its unit price, summed.

An illustrative example (the unit prices are invented for the arithmetic; our database holds no price data at all):

Material	Dose	Assumed price per kg	Contribution
Ethanol (solvent)	78.0%	5	3.90
Bergamot oil	8.0%	150	12.00
Dihydromyrcenol	5.0%	30	1.50
Iso E Super	4.0%	38	1.52
Linalool	3.0%	25	0.75
Patchouli oil	1.5%	200	3.00
Orange blossom absolute	0.5%	7,500	37.50
		Cost per kg	60.17

The absolute costs 50 times what the bergamot costs, but at 0.5% it accounts for 62% of the total. Cut it and you save 37.50, at the price of losing what the formula is about.

The reverse is more interesting. Raise the absolute from 0.5% to 1.5%, taking the extra 1% out of the ethanol, and the cost goes from 60.17 to 135.12 — up 125%. Small changes to expensive materials move the total far more than large changes to cheap ones.

Three practical conclusions:

1. Read the contributions. List "dose × price" for every material and sort it. Usually two or three lines decide the total.

2. To cut cost, move those two or three and leave the rest alone. Taking dihydromyrcenol from 5% to 3% saves 0.60 and may destroy the structure. I have made that trade, and spent more evenings repairing the accord than the saving could ever justify.

3. A cheap material used in bulk becomes a major cost. The ethanol above has the lowest unit price and the third-largest contribution. Always shop around for your bases.

Why a natural material's price can double overnight

The second shock: last year's quotation and this year's, for the same material, can differ by a multiple.

In 2022 Khan, Su, Khurshid and Umar published an analysis in Agricultural and Food Economics applying the GSADF method to quarterly vanilla prices from 1971 to 2020, testing for price bubbles.

The results show that there are five bubbles in the vanilla price that are driven by specific reasons.

The drivers they identify include cyclone damage to growing regions, cartel monopoly control of exports, government deregulation together with export taxation as high as 82%, and later rising demand and price speculation. The most recent bubble runs from the third quarter of 2014 to the third quarter of 2018.

The point is not vanilla. The point is the mechanism the study exposes: the price of a natural material depends on the weather in one growing region, that region's politics, its trade structure, and speculation. None of these has anything to do with your formula, and none of them can be forecast.

The implication: if your cost structure leans heavily on one natural material, your costs are tied to another country's weather. The batch variation article reached the same conclusion (do not let one natural be the only thing holding a formula up) through quality; this article reaches it through money.

Some materials are not a price problem but a supply problem

Then there is a category that money does not reliably solve.

In 2024 Yan and colleagues reviewed the properties of sandalwood oil and the microbial synthesis of its major sesquiterpenoids in Biomolecules. A figure they cite tells the story:

Records show that the annual consumption of Indian sandalwood is approximately 4,000 tonnes. According to 2004 records, oil production has decreased drastically to around 70 tonnes compared to the last six decades, when output was roughly 180 to 200 tonnes.

The subject of that review is itself part of the answer: the whole paper is about engineering yeast and bacteria into cell factories producing santalene and santalol. When natural supply cannot meet demand, the industry goes looking for a synthetic or biosynthetic route. The batch variation article argued that the real natural/synthetic difference is variance; supply is the second difference. A synthetic route can scale, and a tree that takes decades to mature cannot.

Availability in our data

What this section measures. Our database has no price field, so what follows is availability, not cost. The 137,642 supplier records say "this company lists this product" — not that it is in stock today, and not at what price.

After removing two directory-type entries (in practice they are indexes), 21,494 materials carry supplier records.

The first number deserves a pause:

28% (5,897 materials) have exactly one supplier in the entire database.

A single supplier means that if they discontinue it, reformulate, or simply stop answering your email, that material leaves your palette. The batch article asked you to record batch numbers; the reason here is more basic. There may not be a next batch.

But the real threshold is who will sell you a small quantity

Supplier count badly overstates your access, because suppliers differ in kind. Among the largest by record count are research chemical catalogues, selling one-gram vials of research-grade material with a certificate of analysis but no fragrance specification, priced by the milligram. Being on the list does not mean you can buy a usable amount of usable material.

We identified 8 suppliers in the list that clearly serve perfumers in small quantities, and counted how many materials at least one of them carries:

	Count	Share
Materials with any supplier record	21,494	—
Carried by at least one small-quantity retailer	1,743	8.1%
Materials with an odour description	5,824	—
Carried by at least one small-quantity retailer	1,500	26%

The supplier and odour counts above are as of August 2026 and move as the source is re-read; the retailer counts below them are not recomputed with it.

These two figures are lower bounds, not estimates. We counted only the 8 small-quantity retailers we could positively identify in the supplier list; other small-quantity suppliers exist that this database does not include. The true share is higher than 8.1% and 26% — but nowhere near 100%.

The variation between families is large:

Odour family	Available in small quantity
Woody	44.4%
Floral	37.2%
Balsamic	34.4%
Citrus	27.9%
Herbal	25.9%
Spicy	23.7%
Minty	23.6%
Green	22.1%
Fruity	18.0%
Sulfurous	2.8%

Sulfurous sits at 2.8%. That independently confirms a finding from the IFRA article, where the IFRA-versus-FEMA coverage showed the sulfurous, nutty and caramellic end of the database to be mostly flavour territory. Here is the same fact seen from the purchasing side: small-quantity retailers do not stock these materials, because perfumers do not use them.

So if you dig up an intriguing sulfurous material and cannot find it for sale anywhere, your search technique is fine. The material was never in the perfumery market to begin with.

The real barriers to buying small

Four layers stand between a beginner and a material:

1. Minimum order quantity. Trade suppliers normally start at a kilogram and expect a company account. Their packaging and logistics are built at that unit, and hobby quantities simply fall outside the machinery.

2. A B2B account. Many require business registration and a tax number, and some ask what you will use the material for and at what annual volume.

3. Hazardous goods shipping. Ethanol, some aldehydes and various essential oils are flammable: restricted by air, surcharged by road, and worse across borders. This is what routinely turns "the material costs 25" into "it arrives for 80". My own first order did exactly that, and I stared at the freight line on the invoice for a good while.

4. Sample policy. Some suppliers send samples, but usually to account holders.

In practice, build your palette from what you can re-buy from those eight, or their equivalents where you live. The first article said to skip the kit and buy ten single materials; here is the added condition: they should be ten you can still buy in six months. A formula built on a material you cannot buy twice is a work that cannot be reproduced, which is the batch article's conclusion in a more absolute form.

What a purchase record should contain

The evaluation note from the record-keeping article has material, supplier and batch. Add four purchasing fields; every one of them earns its place when you have to remake a formula:

Field	Why
Unit price and unit (per kg or per 100 g)	Without a common unit you cannot compare quotes or compute cost
Minimum order and lead time	Decides whether you can restock when you need to
Total paid including shipping and duty	Hazmat freight frequently exceeds the material
Date of the quotation	Natural material quotes expire; next year is a different price (see vanilla above)

And one habit: recost your main formulae every six months. The saving is incidental. What you are really watching for is which material's price is moving — usually the first signal that supply is in trouble, and you would rather know before the stock-out than after.

---

Unit price decides whether you can bring yourself to buy it, dose decides your cost, and availability decides whether your formula can exist at all. Beginners stare at the first. The third is what kills formulae.

References

K. Khan, C.-W. Su, A. Khurshid & M. Umar, Are there bubbles in the vanilla price?, Agricultural and Food Economics, 10(1), 6 (2022). PMID 35399815

X. Yan et al., Biological properties of sandalwood oil and microbial synthesis of its major sesquiterpenoids, Biomolecules, 14(8), 971 (2024). PMID 39199359

---

Next up is storage: how materials degrade, what oxidation smells like, whether refrigeration and decanting actually help, and how long an opened bottle remains usable.