From the formulation to the optimization of skin creams: A case study based on the Ideal Profile...
-
Upload
lizette-cadman -
Category
Documents
-
view
213 -
download
1
Transcript of From the formulation to the optimization of skin creams: A case study based on the Ideal Profile...
![Page 1: From the formulation to the optimization of skin creams: A case study based on the Ideal Profile Method T. Worch, S. Lê, P. Punter & J. Pagès (thierry@qistatistics.co.uk)](https://reader036.fdocuments.us/reader036/viewer/2022081602/551b2dae55034607418b61aa/html5/thumbnails/1.jpg)
From the formulation to the optimization of skin creams:
A case study based on the Ideal Profile Method
T. Worch, S. Lê, P. Punter & J. Pagès([email protected])
![Page 2: From the formulation to the optimization of skin creams: A case study based on the Ideal Profile Method T. Worch, S. Lê, P. Punter & J. Pagès (thierry@qistatistics.co.uk)](https://reader036.fdocuments.us/reader036/viewer/2022081602/551b2dae55034607418b61aa/html5/thumbnails/2.jpg)
In the literature, many studies have shown that consumers are somewhat accurate when rating products.Still, when consumers are used, another question could be asked…
© Qi Statistics Ltd Slide 2
Introduction
Crunchy, sweet, a lil’ hint of salt…
Hmmm!Delicious!
Aww! No!! Too
sour…
Sensory task
Hedonic taskJAR task
![Page 3: From the formulation to the optimization of skin creams: A case study based on the Ideal Profile Method T. Worch, S. Lê, P. Punter & J. Pagès (thierry@qistatistics.co.uk)](https://reader036.fdocuments.us/reader036/viewer/2022081602/551b2dae55034607418b61aa/html5/thumbnails/3.jpg)
© Qi Statistics Ltd Slide 3
The essence of the IPM
What is your ideal?
![Page 4: From the formulation to the optimization of skin creams: A case study based on the Ideal Profile Method T. Worch, S. Lê, P. Punter & J. Pagès (thierry@qistatistics.co.uk)](https://reader036.fdocuments.us/reader036/viewer/2022081602/551b2dae55034607418b61aa/html5/thumbnails/4.jpg)
Consumers rate products on:– both their perceived and ideal intensities for a list of attributes
• alternatively• using the same scale
– overall liking (and specific liking).
It is a mix of:– quantitative descriptive analysis (such as QDA®) but
• performed by consumers
– JAR scaling but• perceived and ideal intensities are rated explicitly
© Qi Statistics Ltd Slide 4
Principle of the IPM [1]
![Page 5: From the formulation to the optimization of skin creams: A case study based on the Ideal Profile Method T. Worch, S. Lê, P. Punter & J. Pagès (thierry@qistatistics.co.uk)](https://reader036.fdocuments.us/reader036/viewer/2022081602/551b2dae55034607418b61aa/html5/thumbnails/5.jpg)
© Qi Statistics Ltd Slide 5
IPM in practice
![Page 6: From the formulation to the optimization of skin creams: A case study based on the Ideal Profile Method T. Worch, S. Lê, P. Punter & J. Pagès (thierry@qistatistics.co.uk)](https://reader036.fdocuments.us/reader036/viewer/2022081602/551b2dae55034607418b61aa/html5/thumbnails/6.jpg)
© Qi Statistics Ltd Slide 6
IPM Data
P denotes the number of productsA denotes the number of attributes
HEDO
PProducts
Sensory Profile
P Products
A Attributes
Ideal Profile
PProducts
A Attributes
![Page 7: From the formulation to the optimization of skin creams: A case study based on the Ideal Profile Method T. Worch, S. Lê, P. Punter & J. Pagès (thierry@qistatistics.co.uk)](https://reader036.fdocuments.us/reader036/viewer/2022081602/551b2dae55034607418b61aa/html5/thumbnails/7.jpg)
Questions:
What is ideal data?What can we expect from it?
![Page 8: From the formulation to the optimization of skin creams: A case study based on the Ideal Profile Method T. Worch, S. Lê, P. Punter & J. Pagès (thierry@qistatistics.co.uk)](https://reader036.fdocuments.us/reader036/viewer/2022081602/551b2dae55034607418b61aa/html5/thumbnails/8.jpg)
“A conception of something in its absolute perfection.”“One that is regarded as a standard or model of perfection or excellence.”“An ultimate object of endeavor; a goal.”“Considered the best of its kind.”“Completely or highly satisfactory.”
Source: Oxford dictionary
© Qi Statistics Ltd Slide 8
Definition of ideal
![Page 9: From the formulation to the optimization of skin creams: A case study based on the Ideal Profile Method T. Worch, S. Lê, P. Punter & J. Pagès (thierry@qistatistics.co.uk)](https://reader036.fdocuments.us/reader036/viewer/2022081602/551b2dae55034607418b61aa/html5/thumbnails/9.jpg)
Proposition:
“It is a product with particular sensory characteristics
which would maximize liking”
© Qi Statistics Ltd Slide 9
Sensory definition of ideal
SENSORY ASPECT
HEDONIC ASPECT
![Page 10: From the formulation to the optimization of skin creams: A case study based on the Ideal Profile Method T. Worch, S. Lê, P. Punter & J. Pagès (thierry@qistatistics.co.uk)](https://reader036.fdocuments.us/reader036/viewer/2022081602/551b2dae55034607418b61aa/html5/thumbnails/10.jpg)
© Qi Statistics Ltd Slide 10
Expected ideal
![Page 11: From the formulation to the optimization of skin creams: A case study based on the Ideal Profile Method T. Worch, S. Lê, P. Punter & J. Pagès (thierry@qistatistics.co.uk)](https://reader036.fdocuments.us/reader036/viewer/2022081602/551b2dae55034607418b61aa/html5/thumbnails/11.jpg)
Question:
How confident can I be about ideal data?
![Page 12: From the formulation to the optimization of skin creams: A case study based on the Ideal Profile Method T. Worch, S. Lê, P. Punter & J. Pagès (thierry@qistatistics.co.uk)](https://reader036.fdocuments.us/reader036/viewer/2022081602/551b2dae55034607418b61aa/html5/thumbnails/12.jpg)
Sensory consistency:
The ideal data provided by a consumer is consistent if the sensory profile associated to this ideal has similar sensory characteristics as the most appreciated product.E.g.: Consumers who said they have a higher appreciation for the products perceived as sweeter should also rate their ideals as rather sweet.
Statistical procedure:Ensure that the ideal is making the link between sensory and hedonic.
© Qi Statistics Ltd Slide 12
Consistency [2]
![Page 13: From the formulation to the optimization of skin creams: A case study based on the Ideal Profile Method T. Worch, S. Lê, P. Punter & J. Pagès (thierry@qistatistics.co.uk)](https://reader036.fdocuments.us/reader036/viewer/2022081602/551b2dae55034607418b61aa/html5/thumbnails/13.jpg)
Hedonic consistency:
The ideal data provided by the consumers should correspond to products which would be more appreciated than the products tested.The ideal product of a consumer should hence be associated to a liking score (liking potential) which should be larger than the liking scores given to the products tested.
Statistical procedure:Define a model which explains liking based on the sensory and apply this model to the ideal product.
© Qi Statistics Ltd Slide 13
Consistency [3]
![Page 14: From the formulation to the optimization of skin creams: A case study based on the Ideal Profile Method T. Worch, S. Lê, P. Punter & J. Pagès (thierry@qistatistics.co.uk)](https://reader036.fdocuments.us/reader036/viewer/2022081602/551b2dae55034607418b61aa/html5/thumbnails/14.jpg)
Questions:
Why acquiring such data?What is the practical purpose of it?
![Page 15: From the formulation to the optimization of skin creams: A case study based on the Ideal Profile Method T. Worch, S. Lê, P. Punter & J. Pagès (thierry@qistatistics.co.uk)](https://reader036.fdocuments.us/reader036/viewer/2022081602/551b2dae55034607418b61aa/html5/thumbnails/15.jpg)
Optimisation:
The Ideal product should be used to improve the products.The guidance provided should take into consideration whether attributes are drivers of liking or not, as well as the deviation between perceived and ideal intensities.
Statistical procedure:Consider both the deviation from ideal for each attribute and the weight of each attribute on liking.
© Qi Statistics Ltd Slide 15
Tool for optimisation [4]
![Page 16: From the formulation to the optimization of skin creams: A case study based on the Ideal Profile Method T. Worch, S. Lê, P. Punter & J. Pagès (thierry@qistatistics.co.uk)](https://reader036.fdocuments.us/reader036/viewer/2022081602/551b2dae55034607418b61aa/html5/thumbnails/16.jpg)
Questions:
Is there one ideal for all?Would the optimisation be the same
for every consumer/product?
![Page 17: From the formulation to the optimization of skin creams: A case study based on the Ideal Profile Method T. Worch, S. Lê, P. Punter & J. Pagès (thierry@qistatistics.co.uk)](https://reader036.fdocuments.us/reader036/viewer/2022081602/551b2dae55034607418b61aa/html5/thumbnails/17.jpg)
Clustering:
Like for any optimisation procedure, one needs to make sure that homogeneous segments of consumers are defined (in terms of liking). Otherwise, the optimisation procedure might be based on inappropriate product that nobody would like.
Statistical procedure:Any classical/usual clustering technique can be used.
© Qi Statistics Ltd Slide 17
Homogeneous group…of consumers
![Page 18: From the formulation to the optimization of skin creams: A case study based on the Ideal Profile Method T. Worch, S. Lê, P. Punter & J. Pagès (thierry@qistatistics.co.uk)](https://reader036.fdocuments.us/reader036/viewer/2022081602/551b2dae55034607418b61aa/html5/thumbnails/18.jpg)
Single vs. Multiple ideals:
During the test according to the IPM, consumers were asked to describe their ideals based on each product tested. For homogeneous product categories, the consumers should associate the set of products to one unique ideal. If different sub-categories exist, multiple ideals are found.
Statistical procedure:Use the variability between products of the ideal ratings to determine whether a systematic shift is observed across consumers.
© Qi Statistics Ltd Slide 18
Homogeneous group…of products [5]
![Page 19: From the formulation to the optimization of skin creams: A case study based on the Ideal Profile Method T. Worch, S. Lê, P. Punter & J. Pagès (thierry@qistatistics.co.uk)](https://reader036.fdocuments.us/reader036/viewer/2022081602/551b2dae55034607418b61aa/html5/thumbnails/19.jpg)
Ideal of Reference:
Define the sensory profile of the ideal product used as reference. Two points of view can be adopted: either we try to slightly satisfy the majority of consumers (means), or we try to fully satisfy a lower proportion of consumers (IdMap).
Statistical procedure:IdMap: Project on the sensory product space the ideal area of each consumer and define the area that is shared by the majority of consumers. The sensory profile of the ideal product of this area defines the product of reference.
© Qi Statistics Ltd Slide 19
Defining the reference [6]
![Page 20: From the formulation to the optimization of skin creams: A case study based on the Ideal Profile Method T. Worch, S. Lê, P. Punter & J. Pagès (thierry@qistatistics.co.uk)](https://reader036.fdocuments.us/reader036/viewer/2022081602/551b2dae55034607418b61aa/html5/thumbnails/20.jpg)
Finally, the procedure to analyse ideal data (called Ideal Profile Analysis) is done in four steps:
© Qi Statistics Ltd Slide 20
The Ideal Profile Analysis [7]
![Page 21: From the formulation to the optimization of skin creams: A case study based on the Ideal Profile Method T. Worch, S. Lê, P. Punter & J. Pagès (thierry@qistatistics.co.uk)](https://reader036.fdocuments.us/reader036/viewer/2022081602/551b2dae55034607418b61aa/html5/thumbnails/21.jpg)
Example: The Skin Creams study
![Page 22: From the formulation to the optimization of skin creams: A case study based on the Ideal Profile Method T. Worch, S. Lê, P. Punter & J. Pagès (thierry@qistatistics.co.uk)](https://reader036.fdocuments.us/reader036/viewer/2022081602/551b2dae55034607418b61aa/html5/thumbnails/22.jpg)
The products:18 skin creams were formulated8 out of the 18 skin creams were selected (Napping®)
The consumers:72 women from Agrocampus-Ouest (Rennes)
The questions:13 sensory attributes3 specific liking questions + overall liking
© Qi Statistics Ltd Slide 22
Skin Creams
![Page 23: From the formulation to the optimization of skin creams: A case study based on the Ideal Profile Method T. Worch, S. Lê, P. Punter & J. Pagès (thierry@qistatistics.co.uk)](https://reader036.fdocuments.us/reader036/viewer/2022081602/551b2dae55034607418b61aa/html5/thumbnails/23.jpg)
© Qi Statistics Ltd Slide 23
Product Space
EASY TO APPLY ON THE SKIN
PERCEPTION OF FAT
![Page 24: From the formulation to the optimization of skin creams: A case study based on the Ideal Profile Method T. Worch, S. Lê, P. Punter & J. Pagès (thierry@qistatistics.co.uk)](https://reader036.fdocuments.us/reader036/viewer/2022081602/551b2dae55034607418b61aa/html5/thumbnails/24.jpg)
• ConsistencyThe ideal from consumers are consistent both from a sensory and hedonic point of view.
• SegmentationThe consumers were in strong agreement in terms of liking: no segmentation is observed.
© Qi Statistics Ltd Slide 24
Ideal Profile Analysis
![Page 25: From the formulation to the optimization of skin creams: A case study based on the Ideal Profile Method T. Worch, S. Lê, P. Punter & J. Pagès (thierry@qistatistics.co.uk)](https://reader036.fdocuments.us/reader036/viewer/2022081602/551b2dae55034607418b61aa/html5/thumbnails/25.jpg)
© Qi Statistics Ltd Slide 25
Single vs. Multiple Ideal
-4 -2 0 2 4
-4-2
02
4
Single vs. Multiple Ideal
Dim 1 (69.14%)
Dim
2 (
14
.59
%)
1
3
5
8
13
14
151713
5
8
1314
1517
![Page 26: From the formulation to the optimization of skin creams: A case study based on the Ideal Profile Method T. Worch, S. Lê, P. Punter & J. Pagès (thierry@qistatistics.co.uk)](https://reader036.fdocuments.us/reader036/viewer/2022081602/551b2dae55034607418b61aa/html5/thumbnails/26.jpg)
© Qi Statistics Ltd Slide 26
Ideal Mapping
-6 -4 -2 0 2 4 6
-4-2
02
4
Ideal Mapping
Dim1 (69.14%)
Dim
2 (
14
.59
%)
10
10
10
1
3
5
8
13
14
1517
-4-2
02
4
-15 -10 -5 0 5
-10
-50
5
Dim 1 (69.14%)
Dim
2 (
14
.59
%)
Individual ideal confidence ellipses
10
11
110
12
131141
15116171
18
19 2
20
2122
23
2425
26
27
2829
30
31
310
32
33
3435
3637
38
39
4 40
41
42
43
44
45
46
47
48
49
50
51
510
52
53
54
5556
57
58
59
6 606162
63
64
65
66
67
68
69
7
70
71
72
819
![Page 27: From the formulation to the optimization of skin creams: A case study based on the Ideal Profile Method T. Worch, S. Lê, P. Punter & J. Pagès (thierry@qistatistics.co.uk)](https://reader036.fdocuments.us/reader036/viewer/2022081602/551b2dae55034607418b61aa/html5/thumbnails/27.jpg)
© Qi Statistics Ltd Slide 27
Ideal of Reference
![Page 28: From the formulation to the optimization of skin creams: A case study based on the Ideal Profile Method T. Worch, S. Lê, P. Punter & J. Pagès (thierry@qistatistics.co.uk)](https://reader036.fdocuments.us/reader036/viewer/2022081602/551b2dae55034607418b61aa/html5/thumbnails/28.jpg)
© Qi Statistics Ltd Slide 28
Formulation of new products
Formulation of new products:From the 18 original products, 2 have the characteristics of the ideal (products 2 and 9).
Second test:Use of 6 of the 8 previous products + 2 new products65/72 previous consumers Same methodology of the IPM
![Page 29: From the formulation to the optimization of skin creams: A case study based on the Ideal Profile Method T. Worch, S. Lê, P. Punter & J. Pagès (thierry@qistatistics.co.uk)](https://reader036.fdocuments.us/reader036/viewer/2022081602/551b2dae55034607418b61aa/html5/thumbnails/29.jpg)
© Qi Statistics Ltd Slide 29
New Product Space
EASY TO APPLY ON THE SKIN
PERCEPTION OF FAT
![Page 30: From the formulation to the optimization of skin creams: A case study based on the Ideal Profile Method T. Worch, S. Lê, P. Punter & J. Pagès (thierry@qistatistics.co.uk)](https://reader036.fdocuments.us/reader036/viewer/2022081602/551b2dae55034607418b61aa/html5/thumbnails/30.jpg)
1 2 3 5 8 9 15 17
Compact 4.94B 1.55F 5.57A 4.55C 2.06E 3.08D 3.06D 2.97D
Fatty Asp. 4.58AB 1.73D 4.77A 4.31B 2.06D 2.73C 2.80C 2.81C
Yellow 3.97A 1.08E 3.54B 3.16C 0.99E 1.04E 1.59D 1.54D
Shiny 2.45C 4.44A 1.22D 2.89B 4.44A 4.09A 4.09A 4.23A
Smooth 3.35C 3.73BC 3.48C 4.17AB 4.02AB 4.12AB 4.37A 4.35A
Spread 2.51D 5.30A 2.44D 3.44C 5.01AB 4.75B 3.37C 4.66B
Fatty Text. 4.18A 1.81D 4.17A 4.11A 2.31C 2.56C 3.94A 3.06B
Thick 4.25A 1.29E 4.32A 3.78B 1.65DE 1.85D 3.91AB 2.57C
Penetrating 3.20AB 3.55A 3.26AB 3.14AB 3.32AB 3.26AB 2.95B 3.5AB
Film forming 4.12A 3.34B 4.06A 4.10A 2.94BC 2.58C 4.15A 4.18A
Soft 3.53BC 4.20A 3.01D 3.80AB 3.46BCD 3.07CD 3.61B 3.91AB
Fresh 2.27D 3.22A 2.15D 2.48CD 2.89AB 2.88AB 2.64BC 2.67BC
Fatty feel. 3.99A 2.65B 3.83A 3.85A 2.22BC 1.87C 3.97A 3.75A
Liking 2.75C 3.74A 2.34C 3.23B 3.66AB 3.46AB 3.44AB 3.34AB
© Qi Statistics Ltd Slide 30
So, improvements??
![Page 31: From the formulation to the optimization of skin creams: A case study based on the Ideal Profile Method T. Worch, S. Lê, P. Punter & J. Pagès (thierry@qistatistics.co.uk)](https://reader036.fdocuments.us/reader036/viewer/2022081602/551b2dae55034607418b61aa/html5/thumbnails/31.jpg)
© Qi Statistics Ltd Slide 31
So, improvements??
Visual Texture during application
Texture after application Overall liking
Coeff P-value Coeff P-value Coeff P-value Coeff P-value
Product 1 -0,829 0,000 -0,599 0,000 -0,199 0,204 -0,495 0,001
Product 2 -0,012 0,935 -0,187 0,239 0,670 0,000 0,492 0,001
Product 3 -1,194 0,000 -0,503 0,002 -0,664 0,000 -0,904 0,000
Product 5 -0,111 0,470 0,004 0,979 -0,018 0,910 -0,019 0,900
Product 8 0,377 0,014 0,363 0,022 0,430 0,006 0,415 0,007
Product 9 0,620 0,000 0,415 0,009 -0,015 0,926 0,219 0,151
Product 15 0,471 0,002 0,013 0,932 -0,215 0,171 0,193 0,205
Product 17 0,680 0,000 0,493 0,002 0,010 0,949 0,099 0,515
![Page 32: From the formulation to the optimization of skin creams: A case study based on the Ideal Profile Method T. Worch, S. Lê, P. Punter & J. Pagès (thierry@qistatistics.co.uk)](https://reader036.fdocuments.us/reader036/viewer/2022081602/551b2dae55034607418b61aa/html5/thumbnails/32.jpg)
Presentation of the IPM:Sensory methodology;Concepts around the particular data;Statistical methodology around these concepts;
Presentation of a step by step methodology to analyse the data, from the data acquisition to the optimization (passing through quality check and important processing step).
© Qi Statistics Ltd Slide 32
Summary
![Page 33: From the formulation to the optimization of skin creams: A case study based on the Ideal Profile Method T. Worch, S. Lê, P. Punter & J. Pagès (thierry@qistatistics.co.uk)](https://reader036.fdocuments.us/reader036/viewer/2022081602/551b2dae55034607418b61aa/html5/thumbnails/33.jpg)
IPM appears to be a good methodology for product optimisation and product development.
Why using IPM/IPA over other methodologies? all information provided directly from the same consumers; the data collected is rich;
• the quality of the data can be verified;• check for multiple ideals;
extrapolation outside the product space.
© Qi Statistics Ltd Slide 33
Conclusion
![Page 34: From the formulation to the optimization of skin creams: A case study based on the Ideal Profile Method T. Worch, S. Lê, P. Punter & J. Pagès (thierry@qistatistics.co.uk)](https://reader036.fdocuments.us/reader036/viewer/2022081602/551b2dae55034607418b61aa/html5/thumbnails/34.jpg)
[1] Worch, Lê, Punter & Pagès (2013). Ideal Profile Method (IPM): the ins and outs, FQP, 28.
[2] Worch, Lê, Punter & Pagès (2012). Assessment of the consistency of ideal profiles according to non-ideal data for IPM, FQP, 24.
[3] Worch, Lê, Punter & Pagès (2012). Extension of the consistency of the data obtained with the Ideal Profile Method: Would the ideal products be more liked than the tested products? FQP, 26.
[4] Worch, Dooley, Meullenet & Punter (2010). Comparison of PLS dummy variables and Fishbone method to determine optimal product characteristics from ideal profiles, FQP, 21.
[5] Worch & Ennis (2013). Investigating the single ideal assumption using Ideal Profile Method, FQP, 29.
[6] Worch, Lê, Punter & Pagès (2012). Construction of an Ideal Map (IdMap) based on the ideal profiles obtained directly from consumers, FQP, 26.
[7] Worch, Crine, Gruel & Lê (2013). Analysis and validation of the Ideal Profile Method: Application to a skin cream study, FQP, in press.
© Qi Statistics Ltd Slide 34
References
![Page 35: From the formulation to the optimization of skin creams: A case study based on the Ideal Profile Method T. Worch, S. Lê, P. Punter & J. Pagès (thierry@qistatistics.co.uk)](https://reader036.fdocuments.us/reader036/viewer/2022081602/551b2dae55034607418b61aa/html5/thumbnails/35.jpg)
Thank YOU!!
Pieter Punter Sébastien Lê Jérôme Pagès
This presentation is available for download @www.qistatistics.co.uk