Mining the dominant patterns of customer shifts between segments by using top-k and distinguishing sequential rules

Abstract

Purpose – The purpose of this paper is to detect different behavioral groups and the dominant patterns of customer shifts between segments of different values over time. Design/methodology/approach – A new hybrid methodology is presented based on clustering techniques and mining top-k and distinguishing sequential rules. This methodology is implemented on the data of 14,772 subscribers of a mobile phone operator in Tehran, the capital of Iran. The main data include the call detail records and event detail records data that was acquired from the IT department of this operator. Findings – Seven different behavioral groups of customer shifts were identified. These groups and the corresponding top-k rules represent the dominant patterns of customer behavior. The results also explain the relation of customer switching behavior and segment instability, which is an open problem. Practical implications – The findings can be helpful to improve marketing strategies and decision making and for prediction purposes. The obtained rules are relatively easy to interpret and use; this can strengthen the practicality of results. Originality/value – A new hybrid methodology is proposed that systematically extracts the dominant patterns of customer shifts. This paper also offers a new definition and framework for discovering distinguishing sequential rules. Comparing with Markov chain models, this study captures the customer switching behavior in different levels of value through interpretable sequential rules. This is the first study that uses sequential and distinguishing rules in this domain.

A new methodology for customer behavior analysis using time series clustering

“…research studies in Table I denoted by D in the static/ dynamic segmentation). The main drawback of the static segmentation models is that these models assume that customer segments and their members are stable; however, in realworld, customer behavior is not stable and evolves over time (Akhondzadeh-Noughabi and Albadvi, 2015;Böttcher et al, 2009). As mentioned previously (Akhondzadeh-Noughabi and Albadvi, 2015), knowledge and insight gained by exploiting the static models are valid only for a particular time point and do not reflect the complex and evolving nature of customer behavior.…”

“…Reviewing the literature on RFM-based customer segmentation indicates that most of the studies adopted a static segmentation approach. One of the most crucial issues in customer segmentation is emphasizing "customer dynamics" (Akhondzadeh-Noughabi and Albadvi, 2015). In fact, owing to the impact of psychosocial and environmental determinants, including changes of market, promotion, new product introduction and competitors' strategies, customer behavior is changing over time (Akhondzadeh-Noughabi and Albadvi, 2015).…”

“…One of the most crucial issues in customer segmentation is emphasizing "customer dynamics" (Akhondzadeh-Noughabi and Albadvi, 2015). In fact, owing to the impact of psychosocial and environmental determinants, including changes of market, promotion, new product introduction and competitors' strategies, customer behavior is changing over time (Akhondzadeh-Noughabi and Albadvi, 2015). Thus, to be effective, a customer segmentation process must consider the dynamic nature of customer behavior (Chong and Chen, 2010;Böttcher et al, 2009;Ha, 2007;Ha and Bae, 2006;Kim et al, 2005;Chen et al, 2005;Song et al, 2001).…”

“…As pointed out by Akhondzadeh-Noughabi and Albadvi (2015), the main drawback of static segmentation models is that they assume customer segments and their members as stable; however, in the real world, customer behavior is not stable and evolves over time. Furthermore, models achieved using static segmentation approaches are valid only during a particular time point.…”

“…The sequential rule mining and time series analysis are two main techniques which are used in this area. Most of works focused on sequential rule mining (Rahimi et al, 2017;Akhondzadeh-Noughabi and Albadvi, 2015;Böttcher et al, 2009;Song et al, 2001) and there are few papers based on time series clustering in the customer segmentation concept (Saas et al, 2016;Menendez et al, 2014). Moreover, there are two studies that have considered the RFM model on various time points (Song et al, 2017;Hosseini and Shabani, 2015).…”