Trainings & Webinars

Day 1 Exploratory and Modelling methods for Consumer Research

Designed to take you through the key methods that we use in consumer science and explain how they work and how they are applied to consumer data. Emphasis is given to the practical decision making that is made after each analysis. Participants will apply the methods to real life data using XLSTAT routines and will be guided by the written solutions that are a unique feature of Hal MacFie Training. These solutions give you the possibility to pick up the notes after 6 months or a year and remind yourself how a technique works and then apply it to your own data.

Deliverables: Analysis of variance one two and three way applied to sensory and consumer data. Theory, application and analysis experience in Principal Components (with rotation), Correspondence Analysis, Quadratic regression and inspection of response surfaces and the XLSTAT PREFMAP module applied to a single liking variable.

Day 2 Segmentation, Mapping and Portfolio optimisation
Segmentation Masterclass: This will give attendees a unique perspective: understanding and experience with the three main segmentation algorithms. Practical guidance onselecting the correct algorithm for the task, selecting the optimum cluster set, testing for stability will be given.
The key question with clusters is to identify any predominant characteristics of the membership ability and we will show a variety of approaches for this.

Deliverables: AHC, KMEANS and Latent Class Segmentation. Cluster Diagnostics and stability testing. Demographic and Psychographic analyses

Preference Mapping: The workings of the XLSTAT Preference Mapping module will be described, and its application to overall means, segment means and individual scores demonstrated. Product prototype selection and reverse regression to predict the sensory properties of a prototype will be followed by a new section exploring how to integrate the open- ended question responses into the decision making process.. In some categories Preference Mapping on consumer measures are gradually replacing sensory panel scores and a short section will exemplify this approach. Finally an assessment of the role and validity of internal, external and probabilistic models in this application will be discussed

Deliverables: External Preference mapping, Reverse Regression, Using open ended comments, Optimising product portfolios across segments and markets. Comparing internal and external models.

Day 3 What can you do with 3 samples or less?

This section is designed for the very important class of trials that collect CLT or Home Use test on 1 to 3 samples. Preference mapping models are not possible with these trials but there are many techniques that are well suited to product optimisation and can be applied to a single product trial.

Deliverables:Penalty analysis, Ideal Point Profiling, Kano Analysis using CATA data

Concept optimisation and Product matching. Many concept trials ask multiple questions including sensory about one or more concepts. How can we examine segmentation in the response to a single concept? Which concepts do we select for further development?

Deliverables: Factor Analysis and Segmentation on question batteries.

Typically consumer scientists are asked to supply a product that not only matches sensory expectations but enhances the emotional response and reinforces the Brand. We describe PLS on Fit to Concept measures and the use of Multiple Factor Analysis to handle sensory expectation trials. Brand conceptualisations including sensory, emotion and function can be investigated using CATA and Best-worst measures. MAXDIFF scales for prototypes and concept enable the best matching product to be selected.

Deliverables: Partial least Squares, Fit to Concept, Sensory expectation trials, Multiple factor Analysis, Best-Worst, MAXDIFF scales, Concept to product matching.

Hands on Consumer Driven Product Optimization, 3 one-day workshops

Day 1 is designed to take you through the key methods that we use in consumer science and explain how they work and how they are applied to consumer data. Emphasis is given to the practical decision making that is made after each analysis. Participants will apply the methods to real life data using XLSTAT routines and will be guided by the written solutions that are a unique feature of Hal MacFie Training. These solutions give you the possibility to pick up the notes after 6 months or a year and remind yourself how a technique works and then apply it to your own data. Day 2 is a Segmentation Masterclass: This will give attendees a unique perspective: understanding and experience with the three main segmentation algorithms. Practical guidance on selecting the correct algorithm for the task, selecting the optimum cluster set, testing for stability will be given. The key question with clusters is to identify any predominant characteristics of the membership and we will show a variety of approaches for this. Day 3 afternoon Concept optimisation and Product matching. Many concept trials ask multiple questions including sensory about one or more concepts. How can we examine segmentation in the response to a single concept? Which concepts do we select for further development?

Course Outline

Day 1 Exploratory and Analytical methods for Consumer Science

9.00-9.15 Introductions

9.15-10.45 Exploratory- Principal Component Analysis

• Definition, graphical explanation, scaling
• Sample maps, correlation maps, biplots. Interpretation.
• XLSTAT exercises
• Rotating for Interpretability

10.45-11.30 Exploratory- Correspondence Analysis

• Frequency data applications
• Theory and interpretation
• CATA
• XLSTAT analysis

11.30-12.30 Exploratory- Cluster Analysis

• Distance and similarity measures
• Clustering algorithms, dendrogram, defining clusters.
• XLSTAT exercises

12.30 -13.30 Lunch

13.30-14.30 Analytical- Quadratic regression and simple Preference mapping

• Ideal point model, quadratic response, surface response diagrams
• Picking an optimum
• Multi-response problems
• XLSTAT exercise

14.30-15.30 Analytical- Partial Least Squares

• Concept, graphical explanation, cross validation, predictive modelling.
• Liking to sensory, sensory to instrumental.
• Reverse PLS to identify target product.
• XLSTAT exercises

15.45-17.00 Analytical- Path Analysis – Simple and PATHPLSmodelling

• Rotated factor modelling
• XLSTAT exercises
• PATH PLS theory and demonstration

Day 2 Segmentation, Mapping and Product Optimisation

9.00-11.00 Segmentation Masterclass:

• AHC versus Kmeans versus Latent Class Analysis
• Choosing the method. How many clusters?
• PCA versus CVA plotting strategies
• XLSTAT exercises
• Testing the Stability of the solution. How many non-discriminators?
• XLSTAT exercises

11.00-12.00 Demographics, Psychographics and Segments.

• ANOVA approach
• Chi-square, Correspondence and CHAID approach

12.00-13.00 Preference Mapping using the XLSTAT PREFMAP module

• Ideas behind the method
• Different strategies (overall versus segment means or individuals)
• Contour plotting
• Interpretation of output, decision making
• Exercises in XLSTAT

13.00-14.00 Lunch

14.00-15.00 What Sensory Properties will the Desired Product Have?

• Using the preference maps to identify optimal products
• Estimating their sensory properties using reverse regression
• XLSTAT exercise

15.00 -16.00 Using Open ended comments in Preference mapping:

• Role of open ended comments, Quantification
• Alternative Preference Map basis
• XLSTAT exercises

16.00-17.00 Portfolio optimisation and other approaches:

• Selecting products for global markets
• Internal versus External versus Probabilistic models
• A practical approach
• XLSTAT Exercise

Day 3 am What can you do with 3 products or less?

9.00-9.45 JAR Scales and Penalty Analysis

• Analysis of JAR scale data (Friedman v ANOVA)
• Correlation with liking, XLSTAT Penalty Analysis routine
• Interpretation and decision making
• XLSTAT Exercises

9.45-10.45

• Ideal Point Profiling
• Ideal point scoring, Radar Plotting, Analysis, Product Optimisation
• XLSTAT exercise

11.00-12.00

• CATA Data and Kano Impact Analysis
• CATA data, significance testing, Kano approach. 
• XLSTAT CATA routine

12.00-13.00 Lunch

Day 3 pm Concept Optimisation and Product matching

13.00-14.30 Concept and single product trials and analyses

• Data set up and theory
• Factor Analysis and Segmentation analysis of concept question batteries and interpretation
• XLSTAT exercises

14.45-15.45 Matching Product to concept

• PLS of Fit to concept on sensory data
• Expectation, Blind and Informed testing, design, conduct and analysis
• Multiple Factor Analysis theory and XLSTAT Exercises
• Context, Virtual reality trials and analysis

16.00-17.00 Conceptualisation theory, Best-Worst scaling and Analysis

• Matching sensory and emotion to concept.
• XLSTAT exercises

17.00 Close

Anne Hasted

Anne is senior consultant at Qi Statistics Ltd, a UK based consultancy offering a full portfolio of statistical support through training, data analysis, project consultancy and software development. She is a chartered statistician with over 25 years of consultancy experience in a wide range of companies. She has run training workshops worldwide and is recognised for providing “user friendly” training.

Hal Macfie

Dr. Hal MacFie is a statistician by training, with an international reputation in the areas of product assessment and consumer research. He is a Visiting Professor at the Universities of Reading and Nottingham and was for 20 years a senior Editor of Food Quality and Preference.

The quality certification has been delivered for the following category of action:

Training actions