Tool: CONJOINT ANALYSIS

 Concept Testing – is very straightforward

 Executing the concept test

 Positioning concept
 Core idea concept

 Data
to collect - likelihood to buy and
expected number of purchases
 The purpose of conjoint analysis:

 To help in selection of features to be offered in

new product
 To help to set the price of new product
 To predict the resulting level of sales and usages
 To try a new product concept
 To determine the composition of the most
preferred brand
 To determine the relative importance consumer
attach to the salient attributes and the utilities
(part worth)they attach to the levels of
attributes.
 The separate amount of utility provided by each
attribute can be combined

 It works on the simple principal of developing a

part-worth or utility function stating the utility
consumers attach to the levels of each attribute.
Brand Hard Drive RAM Screen Price
Dell 320 GB 2 GB 15.4 in $1,200
Dell 160 GB 4 GB 15.4 in $900
Dell 320 GB 4 GB 12.1 in $1,500
Dell 160 GB 2 GB 12.1 in $1,200
Dell 320 GB 4 GB 12.1 in $900
Dell 160 GB 2 GB 15.4 in $1,500
 Part-worth functions. The part-worth functions, or utility
functions, describe the utility consumers attach to the levels
of each attribute.
 Relative importance weights. The relative importance
weights are estimated and indicate which attributes are
important in influencing consumer choice.
 Attribute levels. The attribute levels denote the values
assumed by the attributes.
 Full profiles. Full profiles, or complete profiles of brands, are
constructed in terms of all the attributes by using the attribute
levels specified by the design.
 Pairwise tables. In pairwise tables, the respondents evaluate
two attributes at a time until all the required pairs of
attributes have been evaluated.
 Determine attributes and attribute levels
 Select product profiles to be measured
 Choose a method of stimulus presentation
 Decide on the response method
 Collect and analyze the data
 Interpret the results
 Identify
the relevant product attributes that are
considered during choice
 Six or seven attributes can be selected
 Attribute and levels should be actionable
 Previous research, secondary data

 Selectattribute levels that represent the options

actually available/feasible in the market
 At least three levels should be used
 The variation in levels must be high
 Suppose the company has conducted a qualitative research and three
attributes, such as sole, upper and price, have been identified as salient
attributes.
 Each attribute is defined in terms of three levels
Level
Attribute Number Description

Sole 1 Rubber
2 Polyurethane
3 Plastic

Upper 1 Leather
2 Canvas
3 Nylon

Price 1 $30.00
2 $60.00
3 $90.00
 Two approaches are used to present the product profile
 Pair wise approach – two attributes at a time

 Full profile approach

 Full profile: complete profile is considered for all

attributes. Each profile is presented on a card. Data_1

 Fractional factorial designs:Data_2

 subset of all possible combinations
 orthogonal designs in which each level of one attribute is
paired equally with all the levels of other attributes
 Sole Sole
Rubber Polyure- Plastic Rubber Polyure- Plastic
thane thane
P
U r $30.00
p Leather
i
p c
e Canvas e $60.00
r

Nylon $90.00

Price
$ 30.00 $60.00 $90.00

Leather
U
p
p Canvas
e
r
Nylon
 Both the methods, that is, pair-wise (two factor) approach and full-
profile approach, have their own utility, but full-profile approach is the
most widely used method.

Example of a Sneaker Product Profile

Sole Made of rubber

Upper Made of nylon
Price $30.00

Full-profile approach for collecting conjoint data

 In the sneaker example, three attributes and three levels of each

attribute are given. Hence, following full-profile approach, a total of 3 ×
3 × 3 = 27 profiles can be constructed.
 Attribute Levels a

Preference
Profile No. Sole Upper Price Rating
1 1 1 1 9
2 1 2 2 7
3 1 3 3 5
4 2 1 2 6
5 2 2 3 5
6 2 3 1 6
7 3 1 3 5
8 3 2 1 7
9 3 3 2 6

a The attribute levels correspond to those in Table 21.2

 Data can be of both the forms: metric data and
non-metric data.

 For non-metric data, the respondents indicate

ranking, and for metric data, the respondents
indicate rating.

 Rating method is most preferred

 the dependent variable is consumer preference

or intention to buy a product (rating or ranking
provided by the customers for buying a product).
 The conjoint analysis model can be represented by the
following formula:
The basic conjoint analysis model may be represented by the
following formula:

m ki
U(X) =  a x ij ij
i =1 j =1
Where:
U(X) = overall utility of an alternative

αij = the part-worth contribution or utility associated with

the j th level (j, j = 1, 2, . . . ki) of the i th attribute
(i, i = 1, 2, . . . m)
xjj = 1 if the j th level of the i th attribute is present
= 0 otherwise
ki = number of levels of attribute i
m = number of attributes
To analyse the conjoint analysis data, dummy variables are
treated as independent or explanatory variables and
preference rating obtained from the respondent is treated as
dependent variable.

Dummy variable is created by k-1 where k is number of

attribute level
For Sole the attribute levels were coded as
follows:
X1 X2
Level 1 1 0
Level 2 0 1
Level 3 0 0
The model estimated may be represented as:
U = b0 + b1X1 + b2X2 + b3X3 + b4X4 + b5X5 + b6X6
Where:
X1, X2 = dummy variables representing Sole
X3, X4 = dummy variables representing Upper
X5, X6 = dummy variables representing Price

For Sole the attribute levels were coded as follows:

whereX1 = 1 if choice is for rubber; 0 otherwise
X2 = 1, if choice is for Polyurethane; 0 otherwise
For Upper:
X3= 1 if choice is for leather;0 otherwise
X4= 1 if choice is for canvas;0 otherwise
For Price:
X5= 1, if choice is for $30; 0 otherwise
X6= 1, if choice is for $60; 0 otherwise
Preference Attributes
Ratings Sole Upper Price
Y X1 X2 X3 X4 X5 X6

9 1 0 1 0 1 0
7 1 0 0 1 0 1
5 1 0 0 0 0 0
6 0 1 1 0 0 1
5 0 1 0 1 0 0
6 0 1 0 0 1 0
5 0 0 1 0 0 0
7 0 0 0 1 1 0
6 0 0 0 0 0 1
The levels of the other attributes were coded similarly. The
parameters were estimated as follows:

b0 = 4.222
b1 = 1.000
b2 = -0.333
b3 = 1.000
b4 = 0.667
b5 = 2.333
b6 = 1.333

Given the dummy variable coding, in which level 3 is the base

level, the coefficients may be related to the part-worths:
 Dummy variable coefficient represents the
difference in part-worth for that level minus
the part worth for base level
a11 - a13 = b1
a12 - a13 = b2

 an additional constraint is necessary since

the part worth is estimated on interval scale
where the origin is arbitrary. The additional
constraint look like:

α11+α12+α13 = 0
 These equations for the first attribute, Sole, are:
α11-α13 = 1.00
α12-α13 = -.333
α11+α12+α13 = 0
• Solving these equations, we get:

α11= 0.778
α12= -0.556
α13= -0.222

 These equations for the second attribute, upper, are:

α21-α23 = 1.00
α22-α13 = .667
α21+α22+α23 = 0

• Solving these equations, we get:

α21= 0.445
α22= 0.111
α23= -0.556
 These equations for the second attribute, Price, are:
α31-α33 = 2.333
α32-α33 = 1.333
α31+α32+α33 = 0

• Solving these equations, we get:

α11= 1.111
α12= 0.111
α13= -1.222
The importance of an attribute can be defined as the range of
part worth αij across the level of that attribute.

Ij = {max(αij) – min (αij) } for each attribute

Sum of ranges = (0.778 - (-0.556)) + (0.445-(-0.556))

of part-worth + (1.111-(-1.222))
= 4.668

The relative importance of an attribute can be calculated by

dividing the importance of that attribute by all attribute
importance

Relative importance of Sole = 1.334/4.668 = 0.286

Relative importance of Upper = 1.001/4.668 = 0.214
Relative importance of Price = 2.333/4.668 = 0.500
 Level
Attribute No. Description Utility Importance

Sole 1 Rubber 0.778

2 Polyurethane -0.556
3 Plastic -0.222 0.286

Upper 1 Leather 0.445

2 Canvas 0.111
3 Nylon -0.556 0.214

Price 1 $30.00 1.111

2 $60.00 0.111
3 $90.00 -1.222 0.500
The goodness of fit of the estimated model should be evaluated. For
example, if dummy variable regression is used, the value of R2 will
indicate the extent to which the model fits the data.
 1
0.8 1.5

0.6 1
0.4
0.2 0.5
Series1
0 0 Series1
-0.2 Rubber Polyurethane Plastic 30$ 60$ 90$
-0.4 -0.5

-0.6 -1
-0.8
-1.5

0.6

0.4

0.2

0
Leather Canvas Nylon Series1
-0.2

-0.4

-0.6

-0.8