engleski

Customer profiles in the antiques and collectibles industry in Croatia using Gaussian Mixture Model clustering: An empirical study

In today’s competitive and ever-changing world, understanding customers’ behavior and needs and categorize customers based on multiple variables is of vital importance. Customer profiling is a critical aspect of customer segmentation that allows SME owners a better understanding of their market and audiences, so that they can better address their marketing efforts to various audiences (clusters) and maximize their effect. The goal of cluster profiling is to achieve a clear vision of who the customer in each cluster is, what the best way to meet his needs is as well as how and what to offer in order to increase sales and customer satisfaction. The main objective of this empirical research was to identify the profile types (»personas«) for a Croatian-based trade and dealer company from the antiques and collectibles industry, based on their demographics, socioeconomics, psychographics, product and style preferences as well as marketing channel preferences, using primary data collected through a survey of 100 existing customers (N=100) of the trade and dealer company in this industry. The study utilizes the XLSTAT software for Gaussian mixture model to cluster customers. The probability of belonging to each cluster is calculated and a classification is usually achieved by assigning each observation to the most likely cluster. The obtained results singled out three clusters (personas): the Successful old-school business owners, the Collector medical doctors and attorneys at law and the Loyal Facebook followers. Discussion and implications focus on a more detailed identification and definition of the personas and thus show SME companies how they could apply this cluster analysis techniques in their business to adjust the offer and the marketing efforts according to the customer profiles in order to better understand and thus serve their customers.

Antiques and collectibles Industry, Customer clustering, Customer profiling, Gaussian mixture model, Republic of Croatia

nije evidentirano