Customer retention and churn prediction in the telecommunication industry: a case study on a Danish university

Abstract

In this study, we explore the possible factors affecting churn in the Danish telecommunication industry and how those factors connect with retention strategies. The Danish telecommunication industry is experiencing a saturated market regarding the number of customers, but the number of service providers has increased significantly in recent years. Due to the high costs of acquiring new customers, the telecommunication industry put great emphasis on retaining customers in such an intensely competitive industry. We employ five machine learning algorithms: random forest, AdaBoost, logistic regression, extreme gradient boosting classifier, and decision tree classifier on four datasets from two geographical regions, Denmark and the USA. The first three datasets are from online repositories, and the last one contains responses from 311 students from Aalborg University collected through a survey. We identify key features extracted by the best-performing algorithms based on five performance measures. Based on that, we aggregate all the features that appear important for each dataset. The results demonstrate that customers’ preferences are not aligned. Among the prominent drivers, we find that service quality, customer satisfaction, offering subscription plan upgrades, and network coverage are unique to the Danish student population. Telecommunication companies need to integrate the sociohistoric milieu of the Nordic countries to tailor their retention policies to different consumer cultures.

Network effects—do they matter for digital technologies diffusion?

“…The macro-effects of technology diffusion (approximated by the share of society having access to given technology) are entirely subjected to the telecommunication companies' market expansion strategy, which is driven mostly by changes in individual preferences and demands. In many works, we find claims, see, for example, works of Geroski (2000) or Peres N uñez and Hilbert (2010), that emerging network externalities are the principal phenomenon driving the process of technology diffusion and determining its in-time dynamics of telecommunication markets (Chen et al, 2022;Saleh and Saha, 2023). The network effects are analogous to the "imitation effect" (Bass, 1980), or the "domino effect" (Lee et al, 2010).…”

Identifying Key Learning Algorithm Parameter of Forward Feature Selection to Integrate with Ensemble Learning for Customer Churn Prediction

“…The strategy of loyal customer retention plays a vital role in a company's growth. Because, only a loyal customer can become a good advocate in a real sense, on the company's behalf, to attract more customers towards it and to increase its profitability [8]. Therefore proactively anticipation about customer churn is inevitable for a Telecommunication Service Provider.…”