Contributions of Clustering Variable Selection: Methods for International Segmentation
Performing international activities is a challenging operation given the heterogeneity of the international market which makes practically impossible the development of successful standardized strategies for the entire world’s population. Finding homogeneous international customer segments helps companies to better communicate with the targeted customers by concentrating on a few units, a group, or several groups. Depending on the study purpose, the segmentation results may help to select potentially attractive international markets, to develop in the context of a global marketing standardized strategies for a segment of countries, or to develop in the context of an international marketing a totally or partially differentiated strategy for several groups. Thus, international segmentation has become an indispensable task in the strategic decision-making process for various international business research questions. Consequently, choosing the relevant segmentation bases and the statistical method represent crucial steps to carry out to identify segments of customers. Actually, research studies in which an international market is segmented mainly employ as bases socio-economic or cultural variables. Moreover, in these studies, since the purpose of the analysis is usually to discover a priori unknown segments in an international population, the segmentation task is performed by clustering techniques. Typically, in this scientific research, to facilitate the interpretation of the results, the segmentation task is preceded by factor analysis to reduce a large number of the initial variables into a few dimensions or factors. However, on the one hand, factor analysis usually generates a loss of information and distortion of reality. On the other hand, the set of the variables initially considered may contain irrelevant variables that might lead to incorrect classification. Therefore, to retain only relevant information for the clustering task: variable selection should be performed to reduce the data dimension before considering a factor analysis.As shown by the numerical experiments,conducted on the basis of two secondary databases: the 03/07/2018 updates of the structure of consumption expenditure published by Eurostat including 32 countries of the European Union and its neighboring countries and the 15/04/2016 version of the updated European Values Study data including customers from 48 European countries, it will allow discovering the accurate groups and facilitate result interpretation. As a result, variable selection allows discovering relevant segments that are easy to interpret. Thus, once the variable selection is performed, the segmentation results will enable relevant and accurate analysis and support correct decision-making.
JEL Classification: C38
Paper type: Empirical research
Copyright (c) 2022 Abdelghafour TALIBI
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.