TY - JOUR
T1 - The polarisation method for merging data files and analysing loyalty to product attributes, prices and brands in revealed preference
AU - Jarvis, Wade
AU - Ridge, C.
AU - Lockshin, L.
PY - 2007
Y1 - 2007
N2 - The method known as revealed preference data is becoming increasingly available for detailed business and academic analysis; however, it is not widely used in describing consumer purchasing beyond the typical statistics reported by the large panel data providers. These statistics are usually related to the brand only and not to other important non-brand attributes. This paper shows how the customer, product and transaction files in one category (wine) are integrated so that a loyalty model can be applied to several product attributes, including prices and brands. The parameters of the model give a simple indication of the level of switching or loyalty that is taking place for any particular attribute and any particular attribute level. The results show that in this category, namely wine, attributes other than the proprietary brand name are driving loyalty. The results also show which specific attribute levels (within an attribute) have higher and lower loyalty. These results are important for marketing practitioners and the paper proposes a rethink in how brands are managed and communicated to consumers in order to optimise performance. When present, variations in attribute-level loyalty require different marketing strategies. For example, high loyalty implies greater deal resistance and more importance on communicating the attribute level, while low loyalty requires constant, small marketing and promotional tactics to maintain market share. As well as a detailed description of the technique, the various marketing implications posited above are drawn out in this paper.
AB - The method known as revealed preference data is becoming increasingly available for detailed business and academic analysis; however, it is not widely used in describing consumer purchasing beyond the typical statistics reported by the large panel data providers. These statistics are usually related to the brand only and not to other important non-brand attributes. This paper shows how the customer, product and transaction files in one category (wine) are integrated so that a loyalty model can be applied to several product attributes, including prices and brands. The parameters of the model give a simple indication of the level of switching or loyalty that is taking place for any particular attribute and any particular attribute level. The results show that in this category, namely wine, attributes other than the proprietary brand name are driving loyalty. The results also show which specific attribute levels (within an attribute) have higher and lower loyalty. These results are important for marketing practitioners and the paper proposes a rethink in how brands are managed and communicated to consumers in order to optimise performance. When present, variations in attribute-level loyalty require different marketing strategies. For example, high loyalty implies greater deal resistance and more importance on communicating the attribute level, while low loyalty requires constant, small marketing and promotional tactics to maintain market share. As well as a detailed description of the technique, the various marketing implications posited above are drawn out in this paper.
M3 - Article
SN - 1470-7853
VL - 49
SP - 489
EP - 513
JO - International Journal of Market Research
JF - International Journal of Market Research
IS - 4
ER -