A1
A2 A3 A6 A5
P
Example Health Drink
A4
Flavor
Nutrition
Solubility Energy
Price
Color
2
Information Processing
Extract a1 = f(A1) Product Information a2 = f(A2) w1a1 w2a2 Product Evaluation Combine
... an = f(An)
Attribute Evaluation
... wnan
Attribute Weighting
Premise - Individuals evaluate the value or utility of a product/service/idea/decision by combining the separate amounts of utility provided by each attribute/factor.
A small sample technique that quantifies peoples preferences or priorities when faced with the task of evaluating a set of products/alternative decisions and choosing the most preferred alternative.
Parallels a realistic purchase/decision situation Often labeled trade-off analysis.
What features or benefits is an individual willing to trade off?
For instance, if you were going to purchase a Health Drink, some of your trade-off considerations might be Flavor Energy Solubility Nutrition Price Color
6
Example
Product category: Health Drink. The attributes and their levels are shown below.
Attributes Levels
Flavor Energy Solubility Color Nutrition Price/kg Chocolate, Elaichi, Plain Low, Medium, High Low, High White, Brown Low, Medium, High 174, 198, 225, 250
8
b) All attributes that potentially create the overall worth should be included.
c) Factors that best differentiate between objects should be included. d) Factors must be distinct and represent a precise concept.
9
2. Selecting Levels
a) Actionable Measures - must be capable of being put into practice. b) Communicable Measures - must be easily communicated for a realistic evaluation.
c) Balanced number of levels - relative importance of a variable increases as the number of levels increases.
d) Set the range of the levels somewhat outside the existing values - not an unbelievable level can reduce inter attribute correlations. Attribute multi-collinearity should be remedied as far as 10 practicable.
2) Interactive Model - In addition to the above, for certain combinations of levels, the total value can be more or less than just their sum. This may decrease predictive power because of reduction in statistical efficiency.
Levels
Chocolate Elaichi Plain Low Medium High Low High White Brown Low Medium High 174 198 225 250 Total
Utility
0.6667 -0.0833 -0.5833 0.6667 -2.3333 1.6667 0.3750 -0.3750 0.0000 0.0000 -0.3333 2.1667 -1.8333 -1.5000 -1.2500 2.5000 0.2500
Importance
1.25*100/14=8.93
Energy
4.00*100/14=28.57
Price/kg
4.00*100/14=28.57
14.00
100
13
Interpreting Part-worths
Using the Parts-Worth values, answer the following questions: 1. What is the most important feature (variable) in the purchase decision for this one individual and what is the preferred level of that feature? 2. What feature is next in importance and what is the preferred level of that feature? 3. What are two features that have less impact on preference?
14
UTILITY
2
LINEAR
2
IDEAL
DISCRETE
15
Part worth diagrams for the attributes of Health Drink for a respondent
0.8 2 0.6 1.5
Estimated utility
0.4
Estimated utility
0.2
0.5
-0.5
-0.2
Chocolate
Elaichi
Plain
Low
Medium
High
-1
-0.4
-1.5
-0.6
-2
-2.5 -0.8
Flavor
0.5
-3
Energy
1
0.4 0.9
Estimated utility
Estimated utility
0.3
0.2
0.1
-0.1
Low
High
-0.2
-0.3
-0.4
-0.5
Brown
2.5
Estimated utility
Estimated utility
1.5
1.5
0.5
0.5
-0.5
Low
Medium
High
-1
-0.5
Rs 174
Rs 198
Rs 225
Rs 250
-1.5
-1
-2
-1.5
-2.5
-2
Nutrition
Price per kg
16
Attribute Importance
30 25
Importance
20
15
10
Flavor
Energy
Solubility
Color
Nutrition
Price
Attributes
17
4. Creating a Stimuli
1. Factorial Design - when all combinations are used. Impractical when number of factors and/or levels is large. 2. Fractional Factorial Design - when a subset of stimuli to be used. Use procedure ORTHOPLAN in SPSS to create the stimuli. Number of stimuli to be used depends on the composition rule selected. Additive model assumes only main effects without interactions - so stimuli should be so selected that they are orthogonal to ensure correct estimation of main effects.
18
20
A total of 20 product concepts are generated using fractional factorial design, 16 of which are used for estimation of model parameters and 4 for cross validation.
21
22
5. Presentation Methods
Use procedure PLANCARD in SPSS to create separate cards containing complete product descriptions for each stimulus for evaluation by the respondents. Trade-off Method Full-Profile Method
Flavor
Choco Elaichi Plain
Energy
Flavor Elaichi Energy Low Solubility Low Color Brown Nutrition High Price per kg 174
Major Limitations
1. Information overload respondents focus on a few factors. 2. Order of factor listing may have impact on evaluation. 23
Note
Total no. of parameters in the model = total no. of levels of all attributes no. of attributes + 1 (3+3+2+2+3+4 6+1= 12). So, this is the minimum no. (NC ) of stimuli (concepts) required to estimate all the parameters in the model. However, if you use NC stimuli, there is a problem (?!) You must use more no. of stimuli to overcome this problem. There is no hard and fast rule to determine the no. of stimuli to be used. In practice, approximately 1.5 to 2 times the minimum no. of (NC ) stimuli is used for the purpose of efficient estimation of the parameters, and testing.
24
7. Estimation
Use procedure CONJOINT( A SYNTAX FILE NEEDS TO BE CREATED) in SPSS to perform conjoint analysis produce a part-worth estimate for each level of each attribute. The mathematical model is (for the Health Drink example):
y b0 b1DFL1 b2 DFL2 b3 DEN1 b4 DEN 2 b5 DSL b6 DCL b7 DNU 1 b8 DNU 2 b9 DPR1 b10 DPR2 b11DPR3 e
25
Flavor
U F 2 U F1 b1 ,U F 3 U F1 b2 U F1 U F 2 U F 3 0
Solve for UF1, UF2, UF3.
Note that SPSS Conjoint procedure reports the utilities (as seen earlier).
26
28
Health Drink 2
Flavor Chocolate Energy Low Solubility Low Color Brown Nutrition High Price per kg 225 0.6667 0.6667 0.3750 0.0000 -1.8333 2.5000
Total Utility
1.6251
Total Utility
2.3751
35% 65%
Elaichi Chocolate
Total
4.0002
Assumptions
The market share is static - total demand is same as that of last year. The Health Drinks 1, 2, 3 will represent the entire market.
32
Therefore,
C1 = 26/41 = 0.63 C2 = 74/59 = 1.25
34
Market Share for Health Drink 3 (Calibrated) 32.2*0.63 = 20.29% (due to Health Drink 1) 32.2*1.25 = 40.25% (due to Health Drink 2)
35
37
After the segments are identified, the demographics of the segments can be compared.
38
40