This study aims to develop a novel model to quantify the strength of customer loyalty based on sequential purchasing behaviors within loyalty programs. It also ranks brands to generate a comprehensive ranking list for each customer segment.
This study proposes a network-based model to record customers’ purchasing history, incorporating purchase frequency, time-decay and discount effects. This study also introduces a modified Hyperlink-Induced Topic Search algorithm to rank brands within each customer segment.
This study analyzes the transactional data set of a multi-industry loyalty program to identify future brand choices for each customer segment.
The proposed methodology does not consider point redemption or expenses incurred for a specific brand. The methodology also does not assume any specific distribution for purchasing time or include predictive analysis.
Loyalty program managers can design marketing strategies based on representative transaction sequence networks from customer segments. They can also identify popular or influential brands. Cross-selling strategies can be developed using information about the brands most likely to be purchased subsequently.
To the best of the authors’ knowledge, this is the first study to propose a network-based model to quantify the strength of customer loyalty from sequential purchasing behaviors. This study also introduces a novel methodology for segmenting customers and proposes a modified Hyperlink-Induced Topic Search algorithm to rank brands.
