A cosmetic brand rarely notices essential oil variation on the day the first drum arrives. It notices it a year later, when the new lot of the same lavender, rosemary or bay laurel makes the serum smell slightly flatter, or pushes the same fragrance dosage over a line the formulator did not know existed. Essential oils are agricultural products, and no supplier can make them behave like a synthetic. What a supplier and a brand can do together is define how much variation is acceptable and build a procurement routine that catches anything beyond it.
Why the same botanical name gives a different oil each year
A distilled oil records the season it came from. Rainfall, heat, harvest date, plant maturity and the way the charge was distilled all move the proportions of the constituents that carry aroma and function. The species and plant part on the label can be identical while the balance of individual molecules shifts from one crop to the next. This is ordinary natural variation, and it is different from adulteration or from a mix-up in chemotype, which are separate problems with their own controls.
For purchasing, the consequence is simple: a specification that says only "lavender essential oil" is not a specification. Consistency has to be written down as a reference and a permitted range, and the supply plan has to allow for the years when the crop lands at the edge of that range.
The reference sample: the anchor for every later lot
The most effective consistency tool costs almost nothing. When your formulation is approved, keep a sealed, labelled portion of the exact oil it was approved with, stored cool and dark in a small, nearly full container so little air sits above the liquid. Alongside it keep the GC-MS chromatogram and the sensory notes from the approval.
Every subsequent lot is then compared with that anchor in two ways. A trained person smells it against the reference on a blotter and records differences in top note, body and dry-down. The chromatogram is laid over the reference profile. Where the two disagree, you have an objective starting point for a conversation with the supplier, instead of an argument about whether the oil "used to smell different". It is also worth refreshing the reference deliberately, as a documented decision, if the brand ever approves a new baseline.
Acceptance ranges that separate normal variation from a problem
A range turns "similar" into something a goods-in team can apply. Agree it with the supplier before the first order, using several past batches if they exist, so the band reflects what the crop really does.
| Check | What you compare | What a breach usually signals |
|---|---|---|
| Appearance and odour | Clarity, colour and aroma against the reference, assessed blind | Oxidation, a different harvest stage, contamination |
| Physical constants | Relative density, refractive index and optical rotation against the agreed windows | Dilution, extension, or a different species or chemotype |
| GC-MS markers | Percentage of the main marker constituents against the band | Seasonal shift at the edge, chemotype change, adulteration |
| Documentation | Lot number on the Certificate of Analysis against the container | Wrong paperwork, repacked or mixed material |
The physical constants are standard methods: ISO 279 covers relative density, ISO 280 refractive index and ISO 592 optical rotation for essential oils. For GC profiles, ISO 11024-1 gives general guidance on preparing a chromatographic profile with a capillary column; its profile shows relative proportions of components rather than absolute concentrations, which is exactly what a marker band needs. Sampling matters too: ISO 212 sets general rules for taking a sample, and a sample drawn from the top of an unmixed drum proves little about the drum.
Set the band deliberately. If it is too narrow, good oil is rejected and the team learns to waive the rule. If it is too wide, the profile creeps away from the reference one acceptable batch at a time.