Article Preview
TopIntroduction
In order to retain customers who have already made a purchase, as well as to attract customers with similar features, companies analyze purchase transactions and form segments to personalize the offer. It is often stated in the literature that retaining existing customers gives the company a stable income (Reichheld & Teal, 1996; Kotler & Armstrong, 2006). Using the RFM model (which proportionally divides customers into segments based on how much they spend, how often they buy and when they last bought), companies can effectively identify valuable customers and then develop an effective marketing strategy (Wei, Lin & Wu, 2010). Then, companies can create a targeting plan for specific potential customers, based on the characteristics of the previous valuable cluster (segment). Therefore, given the competitive environment of contemporary companies, it is of particular importance to identify and retain valuable consumers (Chiliya et al., 2009; Mutandwa et al., 2009). By recognizing patterns and learning from historical data, artificial intelligence and machine learning systems can, with great accuracy, extract recommendations for defined segments, offering products and services that will meet the customers’ needs.
Retail companies who grow in size, do not have only one profitable segment, but several, therefore, objective and effective market segmentation should be conducted. Dividing customers based on their previous purchasing behavior is easier using data mining techniques, such as clustering. In addition, machine learning can, by analyzing previous shopping behavior, predict with what certainty a particular customer will respond to an offer (make a purchase) and, in addition, can predict customer lifetime value (Kietzman, Paschen & Treen, 2018).
The aim of this paper is to identify and compare market segments for two companies involved in the sale and distribution of sports equipment (more details in the methods section). Since one company specializes predominantly in performance sports equipment and the other in lifestyle sports equipment, the aim of the paper is to identify similarities and differences between identified segments (and number of segments) and their previous purchasing behavior. Research in this area often integrates the process of clustering and classification, i.e. combine RFM clustering with predictive classification based on customer and product characteristics to target the most valuable customers but also to anticipate new ones (Cheng et al., 2009; Rogić & Kašćelan, 2019). After defining heterogeneous segments, the optimization of online direct campaigns (placed through social media, for example) is done by customizing the message (Dehghani & Tumer, 2015) for all defined segments.
The rest of the paper is structured as follows: The second section gives an overview of related papers. Section three shows the proposed methodology and describes the datasets used in this research, the fourth section presents the empirical test of the case study, followed by discussion of the results, and concluding remarks with recommendations for the future research.