Confrontation of products spaces based on consumers’ spontaneous data and experts’ conventional profile. Application to the sensory evaluation of twelve luxury fragrances Mélanie Cousin, Maëlle Penven, Mathilde Philippe, Marie Toularhoat, Marine Cadoret, Sébastien Lê Agrocampus,Laboratoire de Mathématiques appliquées, 65 Rue de Saint-Brieuc, 35042 Rennes Cedex
INTRODUCTION This study tries to evaluate the consumers’ ability to agree on a consistent products space and compares consumers’ products space to experts’ one. In this context, sensory analysis sessions have been conducted with a products space made of twelve luxury fragrances [1] [2].
CONSUMERS’ DATA: NAPPING 40
EXPERTS’ CONVENTIONAL PROFILE
30
9
20
7
4
10
10
•
2 8
1
11
12
0
6
0
10
20
30
40
0
99 consumers Spontaneous data • 2 fragrances are close if they are percieved as similar and far when they are judged differently •
3 5
50
5
10
12 experts at fragrances 2 sessions • 12 compulsory descriptors • Scaling evaluation •
SPICY
Not spicy
Very spicy
GREEN
Not green
Very green
MARINE NOTES
No marine
Very marine
•
60
METHODOLOGY: HMFA Products Napping Conventional profile 1 X1 Y1 … X99 Y99 Descriptor 1 … Descriptor 12 xik … xij 12 Consumer 1
CONSUMERS’ NAPPING
Consumer 1
EXPERTS’ PROFILE
Consumer 99
The HMFA (Hierarchical Multiple Factor Analysis)[3], by attributing the same weight to the different groups on each level of the hierarchy, helps us to confront the two points of view on the same graph.
Descriptors 1…12
99 groups
xik: mean score for the k descriptor and the i product
xij: coordinate of the i product for the j consumer
1 group
RESULTS 1.0
CONSUMERS EXPERTS
Aromatics Elixir (EP) Pleasures (EP)
1
Chanel n°5
Floral notes
J’adore (EP)
0.5
Green
Citrus fruit
Pure Poison (EP)
Woody Marine notes
Shalimar (ET)
L’instant (EP)
Spicy Oriental notes
0.0
Dim 2 (17.77%)
Coco Mademoiselle (EP)
0
Dim 2 (17.77 %)
J’adore (ET)
Heady Fruity
Cinéma (EP) Surrounding
-0.5
-1
Angel (EP)
-2
Lolita Lempika (EP)
Vanilla -1.0
Greedy
-2
-1
0
1
-1.0
2
-0.5
• The proximity of the partial points of a product conveys a strong consensus between experts and consumers, especially for the fragrances Coco Mademoiselle, Pure Poison and L’Instant.
The products space given by consumers is nearly the same as the one given by experts. •
0.0
0.5
1.0
Dimension 1 (40.83%)
Dim 1 (40.83 %)
• The descriptors used by experts are particularly well correlated with the two first axes, hence a clear interpretation. •
3 groups are visible: - Woody, spicy and oriental fragrances: Aromatics Elixir, Shalimar - Floral, Fresh and citrus fragrances: J’adore, Pleasures - Vanilla and greedy fragrances: Lolita Lempicka
CONCLUSION • Non-trained and inexperienced consumers are able to find an agreement on the description of twelve luxury fragrances by using a very spontaneous data collection method.
Moreover the HMFA highlights that differences made by consumers are the same than those made by the experts, both obtained products spaces distinguish the same three groups.
•
• An interesting extension of this study is the analysis of textual data given by the consumers for the characterisation of the products. References: [1] Gazano G., Ballay S., Eladan N., Sieffermann J.M. (2005). Flash Profile and fragrance research:using the words of the naive consumers to better grasp the perfume's universe In: ESOMAR Fragrance Research Conference, 15-17 May 2005, New York, NY. [2] Gazano G., Ballay S., Sieffermann J.M. (2006). Transposing fragrance olfactory research into the consumers' worlds and words. In: 24th Congress of the International Federation of Societies of Cosmetic Chemists (October 16-19, 2006, Osaka, Japan), PD-162, 9 pp. [3] Le Dien S. & Pagès J. (2003). Hierarchical Multiple Factor Analysis: application to the comparison of sensory profiles. Food Quality and Preference. 14. pp. 397-403.
