SESH FRAGRANCE ACADEMY | ADVANCED SENSORY EVALUATION
How do you turn mixed sensory scores into a defensible go-or-revise decision?
Follow sample introduction, column separation and detection. This GC schematic is not a SESH perfume test and does not independently establish sensory experience.
Open full-size image ↗Produce a traceable professional report with study aim, method, measured outcomes, uncertainty, limitations and actionable decision.
1. Start with the question and the decision it informs
Avoid opening with 'the perfume is wonderful'. Record project ID, formula revision and the precise purpose, for example whether F03 reduced sharpness without losing floral identity. Quote the relevant brief version, predeclared acceptance criteria and the decision owner. The report should support a repeatable go-or-revise decision, not a story assembled after looking at the scores.
2. A concise executive summary with real limitations
Use four components: study design, sample size, key observation and conditional recommendation. For example, 'In two sessions with eight trained assessors, F03 was rated less sharp than F02, but floral identity approached the rejection boundary; we recommend a further blinded replication before selection.' This communicates what was and was not established rather than inflating incomplete analysis.
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. Minimum reproducibility metadata
Include sample codes, batch IDs, preparation and observation times, substrate, applied amount, test environment, assessor selection, statistical approach and deviations. A study comparing a two-day-old strip with a freshly sprayed one is confounded and should disclose that discrepancy. Use coded assessor IDs and appropriate privacy practices; unnecessary personally identifiable data do not belong in the technical output.
4. Results and interpretation should be separate
In Results, record observed data: assessor count, descriptor-score distribution, version differences where appropriate, repeatability and any discrimination-test outcomes. In Discussion, evaluate hypotheses, such as formulation changes versus delivery differences. 'Observers rated less sharpness' is not evidence that a particular molecule disappeared. Explicitly list unmeasured endpoints and competing explanations.
5. Use an acceptance matrix instead of vague applause
Suppose the brief requires sharpness ≤2/5, floral identity ≥3/5 and descriptive repeatability across sessions. Hypothetical F02 scores sharpness 3.8 and floral 4.1, while F03 scores 1.9 and floral 2.7. F03 meets sharpness but fails floral identity in this educational scenario. Record failure of the combined brief, subject to statistical uncertainty rather than choosing a version solely because one defect improved.
6. Three defensible decision outcomes
Close with one of three outcomes: advance for additional technical verification when predefined sensory criteria pass, revise when a key requirement robustly fails, or repeat/investigate when evidence is insufficient or confounded. None of these is a manufacturing or sale authorization. Real products still require safety, stability, packaging and quality approvals. Preserve the reference formula and versioned report trail.
7. Make the next experiment explicit
Replace 'needs improvement' with a testable next step: re-evaluate F03 after one targeted change to the sharpness–floral balance, holding dose, sample age and substrate constant, with blinded observations at five and thirty minutes. Assign internal ownership and a measurable outcome. A high-quality report identifies the next experiment and prevents the team from repeating the same uncontrolled comparison.
Case study: a partial success that should not be approved
F02 scores sharpness 3.8 and floral identity 4.1; F03 scores 1.9 and 2.7. With criteria sharpness ≤2 and floral identity ≥3, the revision fixes one issue but misses another. These are fictional teaching scores; no final product approval follows. Recommend targeted revision and replication rather than a premature 'go'.
Professional practice exercise
Write a one-to-two-page professional report for a fictional F02/F03 comparison. Include a criteria matrix, session method, raw summary, one potential confounder, one of three decision outcomes and a testable next step. Separate observations from hypothesized mechanisms.
Three educational concept maps
An auditable report structure
- Question and predeclared brief
- Samples and protocol
- Results and uncertainty
- Deviations and limitations
- Decision and next experiment
Illustrative acceptance matrix
- F02: sharp 3.8 / floral 4.1
- F03: sharp 1.9 / floral 2.7
- Criteria: sharp≤2, floral≥3
- F03: partial improvement, fails total brief
Three responsible decision states
- Advance: for further verification
- Revise: targeted correction
- Repeat: insufficient evidence
- None equals safety or sales approval
Scientific glossary
- Traceability
- Ability to link findings to formulas, batches, sessions and report revisions.
- Acceptance matrix
- Table comparing each observed outcome with prespecified requirements.
- Go / Revise / Repeat
- Sensory-stage decisions that precede technical qualification.
Official scientific references
This educational explanation does not constitute ISO certification or full implementation of a standard. Consult the official standard for formal testing.
