We think you have liked this presentation. If you wish to download it, please recommend it to your friends in any social system. Share buttons are a little bit lower. Thank you!
Presentation is loading. Please wait.
Published byLizette Cadman
Modified over 2 years ago
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@example.com)
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 LtdSlide 2 Introduction Crunchy, sweet, a lil’ hint of salt… Hmmm! Delicious! Aww! No!! Too sour… Sensory task Hedonic task JAR task
© Qi Statistics LtdSlide 3 The essence of the IPM What is your ideal?
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 LtdSlide 4 Principle of the IPM 
© Qi Statistics LtdSlide 5 IPM in practice
© Qi Statistics LtdSlide 6 IPM Data P denotes the number of products A denotes the number of attributes HEDOHEDO P Products Sensory Profile P Products A Attributes Ideal Profile P Products A Attributes
Questions: What is ideal data? What can we expect from it?
“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 LtdSlide 8 Definition of ideal
Proposition: “It is a product with particular sensory characteristics which would maximize liking” © Qi Statistics LtdSlide 9 Sensory definition of ideal SENSORY ASPECT HEDONIC ASPECT
© Qi Statistics LtdSlide 10 Expected ideal
Question: How confident can I be about ideal data?
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 LtdSlide 12 Consistency 
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 LtdSlide 13 Consistency 
Questions: Why acquiring such data? What is the practical purpose of it?
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 LtdSlide 15 Tool for optimisation 
Questions: Is there one ideal for all? Would the optimisation be the same for every consumer/product?
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 LtdSlide 17 Homogeneous group… of consumers
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 LtdSlide 18 Homogeneous group… of products 
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 LtdSlide 19 Defining the reference 
Finally, the procedure to analyse ideal data (called Ideal Profile Analysis) is done in four steps: © Qi Statistics LtdSlide 20 The Ideal Profile Analysis 
Example: The Skin Creams study
The products: 18 skin creams were formulated 8 out of the 18 skin creams were selected (Napping®) The consumers: 72 women from Agrocampus-Ouest (Rennes) The questions: 13 sensory attributes 3 specific liking questions + overall liking © Qi Statistics LtdSlide 22 Skin Creams
© Qi Statistics LtdSlide 23 Product Space EASY TO APPLY ON THE SKIN PERCEPTION OF FAT
Consistency The ideal from consumers are consistent both from a sensory and hedonic point of view. Segmentation The consumers were in strong agreement in terms of liking: no segmentation is observed. © Qi Statistics LtdSlide 24 Ideal Profile Analysis
© Qi Statistics LtdSlide 25 Single vs. Multiple Ideal
© Qi Statistics LtdSlide 26 Ideal Mapping
© Qi Statistics LtdSlide 27 Ideal of Reference
© Qi Statistics LtdSlide 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 products 65/72 previous consumers Same methodology of the IPM
© Qi Statistics LtdSlide 29 New Product Space EASY TO APPLY ON THE SKIN PERCEPTION OF FAT
1235891517 Compact 4.94 B 1.55 F 5.57 A 4.55 C 2.06 E 3.08 D 3.06 D 2.97 D Fatty Asp. 4.58 AB 1.73 D 4.77 A 4.31 B 2.06 D 2.73 C 2.80 C 2.81 C Yellow 3.97 A 1.08 E 3.54 B 3.16 C 0.99 E 1.04 E 1.59 D 1.54 D Shiny 2.45 C 4.44 A 1.22 D 2.89 B 4.44 A 4.09 A 4.23 A Smooth 3.35 C 3.73 BC 3.48 C 4.17 AB 4.02 AB 4.12 AB 4.37 A 4.35 A Spread 2.51 D 5.30 A 2.44 D 3.44 C 5.01 AB 4.75 B 3.37 C 4.66 B Fatty Text. 4.18 A 1.81 D 4.17 A 4.11 A 2.31 C 2.56 C 3.94 A 3.06 B Thick 4.25 A 1.29 E 4.32 A 3.78 B 1.65 DE 1.85 D 3.91 AB 2.57 C Penetrating 3.20 AB 3.55 A 3.26 AB 3.14 AB 3.32 AB 3.26 AB 2.95 B 3.5 AB Film forming 4.12 A 3.34 B 4.06 A 4.10 A 2.94 BC 2.58 C 4.15 A 4.18 A Soft 3.53 BC 4.20 A 3.01 D 3.80 AB 3.46 BCD 3.07 CD 3.61 B 3.91 AB Fresh 2.27 D 3.22 A 2.15 D 2.48 CD 2.89 AB 2.88 AB 2.64 BC 2.67 BC Fatty feel. 3.99 A 2.65 B 3.83 A 3.85 A 2.22 BC 1.87 C 3.97 A 3.75 A Liking 2.75 C 3.74 A 2.34 C 3.23 B 3.66 AB 3.46 AB 3.44 AB 3.34 AB © Qi Statistics LtdSlide 30 So, improvements??
© Qi Statistics LtdSlide 31 So, improvements?? Visual Texture during application Texture after application Overall liking CoeffP-valueCoeffP-valueCoeffP-valueCoeffP-value Product 1 -0,8290,000-0,5990,000-0,1990,204-0,4950,001 Product 2 -0,0120,935-0,1870,2390,6700,0000,4920,001 Product 3 -1,1940,000-0,5030,002-0,6640,000-0,9040,000 Product 5 -0,1110,4700,0040,979-0,0180,910-0,0190,900 Product 8 0,3770,0140,3630,0220,4300,0060,4150,007 Product 9 0,6200,0000,4150,009-0,0150,9260,2190,151 Product 15 0,4710,0020,0130,932-0,2150,1710,1930,205 Product 17 0,6800,0000,4930,0020,0100,9490,0990,515
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 LtdSlide 32 Summary
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 LtdSlide 33 Conclusion
 Worch, Lê, Punter & Pagès (2013). Ideal Profile Method (IPM): the ins and outs, FQP, 28.  Worch, Lê, Punter & Pagès (2012). Assessment of the consistency of ideal profiles according to non-ideal data for IPM, FQP, 24.  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.  Worch, Dooley, Meullenet & Punter (2010). Comparison of PLS dummy variables and Fishbone method to determine optimal product characteristics from ideal profiles, FQP, 21.  Worch & Ennis (2013). Investigating the single ideal assumption using Ideal Profile Method, FQP, 29.  Worch, Lê, Punter & Pagès (2012). Construction of an Ideal Map (IdMap) based on the ideal profiles obtained directly from consumers, FQP, 26.  Worch, Crine, Gruel & Lê (2013). Analysis and validation of the Ideal Profile Method: Application to a skin cream study, FQP, in press. © Qi Statistics LtdSlide 34 References
Thank YOU!! Pieter PunterSébastien LêJérôme Pagès This presentation is available for download @ www.qistatistics.co.uk
The basics for simulations
Static Equilibrium; Elasticity and Fracture
Break Time Remaining 10:00.
Clock will move after 1 minute
AGVISE Laboratories %Zone or Grid Samples – Northwood laboratory
Before Between After.
2011 FRANKLIN COMMUNITY SURVEY YOUTH RISK BEHAVIOR GRADES 9-12 STUDENTS=332.
The 5S numbers game..
2011 WINNISQUAM COMMUNITY SURVEY YOUTH RISK BEHAVIOR GRADES 9-12 STUDENTS=1021.
Resistência dos Materiais, 5ª ed.
Schutzvermerk nach DIN 34 beachten 05/04/15 Seite 1 Training EPAM and CANopen Basic Solution: Password * * Level 1 Level 2 * Level 3 Password2 IP-Adr.
Select a time to count down from the clock above
2011年上半年 我院团学工作活动图片展播 2011年8月28日.
Time and Labor Processing Day 1
A Fractional Order (Proportional and Derivative) Motion Controller Design for A Class of Second-order Systems Center for Self-Organizing Intelligent.
Trend for Precision Soil Testing % Zone or Grid Samples Tested compared to Total Samples.
1 Non Deterministic Automata. 2 Alphabet = Nondeterministic Finite Accepter (NFA)
Budapest University of Technology and Economics, BME , 1872
Progressive Aerobic Cardiovascular Endurance Run
INSTALLING "WINDOWS 8.." SO THAT IT CREATES A LOCAL USER ACCOUNT
Copyright © Action Works 2008 All Rights Reserved - Photos by David D
MaK_Full ahead loaded 1 Alarm Page Directory (F11)
CHAPTER 18 The Ankle and Lower Leg
EuroCondens SGB E.
Numeracy Resources for KS2
Add Governors Discretionary (1G) Grants Chapter 6.
: 3 00.
1 Copyright © 2013 Elsevier Inc. All rights reserved. Appendix 01.
TCCI Barometer March “Establishing a reliable tool for monitoring the financial, business and social activity in the Prefecture of Thessaloniki”
Copyright © 2012, Elsevier Inc. All rights Reserved. 1 Chapter 7 Modeling Structure with Blocks.
Introduction The 2014 Report contains an analysis of 1,483 cases of occupational fraud that occurred in more than 100 countries. Figure 1 illustrates the.
Introduction Embedded Universal Tools and Online Features 2.
Addition and Subtraction Equations
Lial/Hungerford/Holcomb/Mullins: Mathematics with Applications 11e Finite Mathematics with Applications 11e Copyright ©2015 Pearson Education, Inc. All.
1 Prediction of electrical energy by photovoltaic devices in urban situations By. R.C. Ott July 2011.
When you see… Find the zeros You think….
Sampling in Marketing Research
突破信息检索壁垒 －SciFinder Scholar 介绍
© 2017 SlidePlayer.com Inc. All rights reserved.