Tuesday, July 8, 2008

Attribute based perceptual mapping by using discriminant analysis for different brands of baby stores in UAE

Dr. Rajashekhar Karjagi and Dr. Joson Jose, Research Analysts Landmark Research Centre, Landmark Group, Dubai

Imagine walking with your family into Baby store on a weekend to pick a toy for your kid. There are 5 top brands of stores fighting for your attention; however you made your choice and walked towards BabyShop. Your family is happy. Thank you!

Now..., that is precisely where my problem starts! Why did you enter that particular store? Was it so jingle, was it the smile of the sales staff, recommended by your colleague, or the store which you went has a better quality of products (simple economics)!

I and my colleague Joson tried to crack the problem. The traditional approach taken to solve the problem were multidimensional scaling or correspondence analysis but we have used discriminant analysis and the results are as discussed below in detail.

Well, we chose to do things differently. Why? We felt that there is another dimension to the decision making which if overlooked can give misleading results! Your sales projections if you are managing any of the brands can go way off the mark.

Here is what we did, we call it the 'Attribute based perceptual mapping by using discriminant analysis'
Methodology
Study area - All the emirates of UAE
Sample size - 1300
Method of data collection - personel interview at the Babyshop stores


Interpretation for Attributes and Dimensions
•It could be observed from the perceptual map in the above figure that, Adams, Mothercare, Toys R us, BabyShop and Others, these five brands (Or Groups) have their unique positions on the map.
• On the same map values of attributes have been plotted on two dimensions (Each discriminant function represents a dimension).
• Dimension 1 (DF1) comprise Product Display, Ambience, Signage, Store Layout, Staff Courtesy, Window Display and Music since the vectors of these attributes are closer to dimension1 – These store attributes contribute more to the dimension 1 and hence this dimension is named as Store Features dimension
• Dimension 2 comprise of only Product Quality which can be evident from the long arrow closer to vertical axis (DF2) and hence its named as Product Quality dimension

Interpretation for Brands and their associations
• BabyShop seem to be stronger in Dimension 1 (Product Display, Ambience, Signage, Store Layout, Staff Courtesy, Window Display and Music), Mothercare and Toys R Us seems to be strong in Dimension 2 (Product Quality). Whereas Adams and Others seems to be weak in both the dimensions as compared to its competitors
• It could be noted that the length of a vector pointing towards a particular brand indicate the association of that attribute with a particular brand and the vectors pointing towards opposite direction from a given brand represent lower association with a particular brand

Significance of the model
•The model was found to be significant at 1 per cent level as could be evident from Box’s F statistic for testing the significance of discriminant analysis.
•The discriminating power (Variation) of Dimension 1 and 2 were found to be 82.9 per cent and 14.3 per cent respectively. It means that Dimension 1 contributes highest to BabyShop to be different from other brands whereas Dimension 2 contributes only 14.3 per cent for discrimination of BabyShop from other brands
•Wilk’s Lambda was found to be significant at 1 per cent level for both the dimensions (DF1 and DF2)
•Although Store Layout and Signage were found to be closer to Dimension 2 they were included under Dimension 1 as these attributes were again in association with BabyShop and no much effect on other brands
•People perceive that BabyShop is known for Store Features rather than product Quality, it means that there is a large gap in the perceived product quality which is very much concern for Babyshop and it can be considered as a weakness as compared with its major competitors (Mothercare and Toys R Us) as the Product Quality is the main strength for them

Conclusion
Although many of the analysts use multidimension scaling technique for analysing the consumer perceptions it is better to go one step ahead by the method of attribute based perceptual mapping by using discriminat analysis or factor analysis but it needs little bit of efforts in data collection and sound knowledge of these techniques.

No comments: