Conjoint Analysis using R programming language to determine customer’s preferences
Conjoint analysis is an effective survey technique to determine customer’s preferences. In this article, a conjoint analysis survey for an RC toy has been designed and tested with dummy customer data. An orthogonal design has been carried out in R programming language and the design has been tested with randomly generated customer’s dummy data set to ensure the reliability of the design.
The conjoint analysis survey evolves six steps:
- Select Attributes
- Determine attribute levels
- Determine attribute combinations
- Select form of presentation of stimuli
The RC toy has the following attributes for which customers’ preferences need to be determine
A. Price: High, Medium, Low
B. Appearance: Cute/funny, Real miniature
C. Size: 1:18 scale, 1:32 scale, 1:48 scale
D. Battery: Rechargeable, Not-rechargeable
Now There are 3*2*3*2 =36 combinations of toy profile and customers have to rank these 36 cards. The task is a tiresome job and it is quite possible that the datasets collected are full of garbage observation. Therefore there is a need to reduce the number of cards in such a way that there is no information loss.
The number of cards can be reduced with orthogonal design. A package named, conjoint, is provided by R, in which such survey can be designed, tested and evaluated.
The product profile by orthogonal design has been shown below. clearly the 36 numbers of product profile has been reduced to 9 numbers of product profile without the loss of information.
A dummy random datasets of 100 customers have been generated and the design has been tested. Following are the results have been obtained. The utility for each customer can be obtained by a regression model.
The average customers’ utility for price, appearance, size, and battery are presented in the following figure.
Now, with the above design presented, a market survey can be conducted to determine the utility of various product’s attributes.