SESH FRAGRANCE ACADEMY | ADVANCED SENSORY EVALUATION
What separates ten casual opinions from a documented sensory panel?
Follow the route from the nasal cavity to the olfactory bulb. Smell is perception; it does not independently identify a perfume’s ingredient list.
Design a descriptive fragrance study with defined objectives, assessor selection, calibration, coded samples and controlled conditions.
1. Begin with one precise research question
You might want a reproducible description of F02, proof of a perceivable difference from F01, or evidence that target consumers prefer F02. These require different designs: trained descriptive assessors and a shared lexicon, a discriminatory forced-choice test, or consumer hedonic research. An internal staff panel can support technical quality decisions but does not automatically represent customers. Document the research question, intended decision and target population before sampling.
2. Select and monitor assessors rather than fans
A descriptive panel needs assessors who can distinguish, scale and consistently apply terms, not participants chosen simply because they love the brand. Screen for suitability, account for temporary illness or strong odor exposure, obtain informed participation and allow withdrawal. Train descriptors against appropriate reference standards, then monitor repeatability across sessions. ISO 8586:2023 addresses assessor selection and training; consumer preference studies require a different sampling logic.
These curves show how changing mean and variance affects a statistical model. They are theoretical teaching data, not skin or fragrance-panel results.
Open full-size image ↗3. Build a calibrated sensory lexicon
Create a small lexicon tied to the brief: citrus, floral, sweetness, powderiness, woodiness, sharpness and perhaps comfort. Define each term with suitable, safe references and clear scale anchors, such as 0 for not perceived and 5 for very strong. Record sensory intensity separately from liking; a panelist can correctly rate woodiness as 5 and still dislike it. ISO 13299:2016 provides guidance for constructing sensory profiles.
4. Control the session before asking opinions
Use an odor-neutral area with documented temperature, ventilation and lighting. Present equivalent, safely prepared samples with controlled dose and age on an appropriate substrate; do not automatically apply experimental formulas to skin. Blind-code samples and balance presentation order to limit adaptation and position effects. Document breaks between samples. Coffee beans are not a proven universal reset switch. ISO 8589:2007 provides guidance on sensory test-room design.
5. Repeatability and agreement answer different questions
If one assessor rates coded duplicates at 4 and 4 for citrus, that is evidence of within-assessor repeatability. If others rate the same sample at 1, investigate reference training, dose differences and perceptual variability. Distinguish repeatability within a person from agreement between people. Allow 'not detected' when the task permits; do not pressure panelists to invent notes. Preserve raw observations even when they contradict the expected result.
6. Design a traceable session sheet
A defensible row includes session ID, coded assessor, sample code, batch, formula revision, substrate, preparation amount, observation time, presentation order, descriptor scores and protocol deviations. Keep identifying information separate when possible. Tie all observations to versions and conditions so that any sensory drift can be audited rather than attributed to a change in packaging without evidence.
7. Turn the panel output into a bounded conclusion
A bounded result might be: 'Across two sessions with six trained assessors, F02 showed a higher descriptive floral score at 30 minutes.' It does not prove universal preference or safety. Descriptive assessment cannot replace chemical stability tests, finished-product compliance or consumer liking research. Report panel size, dispersion, reproducibility and deviations before using the results to revise a formula.
Case: one fragrance, two unrelated decisions
Product development asks whether F02 is less sharp than F01; marketing asks whether target consumers like it more. A trained panel with anchored sharpness ratings addresses the first, whereas an appropriately sampled hedonic test addresses the second. Five enthusiastic employees cannot support a universal consumer-preference claim.
Professional practice exercise
Draft a one-page protocol comparing five descriptors for F01 versus F02 using a fictional six-assessor panel, two sessions and one coded duplicate. Define deviation handling and distinguish descriptive data from liking.
Three educational concept maps
The professional panel pipeline
- Define question and decision
- Select panel type
- Train a common lexicon
- Code and balance presentation
- Review reliability and report
Controlled session map
- Odor-controlled environment
- Matched dose and substrate
- Coded samples, balanced order
- Rest to limit adaptation
Three distinct outcomes
- Intensity: descriptor strength
- Discrimination: detectable difference
- Hedonic: preference score
- Safety: requires separate evaluation
Scientific glossary
- ISO 8586
- Guidance for choosing, training and monitoring sensory assessors.
- ISO 13299
- Framework for constructing sensory profiles.
- Panel calibration
- Training with shared descriptors and anchors to reduce interpretive variation.
Official scientific references
This educational explanation does not constitute ISO certification or full implementation of a standard. Consult the official standard for formal testing.
