AbstractSince dawn of time, merchants try to know their best customers and especially to meet their expectations. More recently, since the early 1980s, managers have aimed to replicate the same strategy using information systems as human judgement cannot efficiently deal with a large number of customers. This approach is called Customer Relationship Management (CRM). In particular, analytical CRM is dedicated to finding commercial opportunities inside customer data.
Traditionally, research on this topic investigates new data mining techniques to improve this identification of opportunities. A more recent approach consists in finding new data or new ways to use the data. In this context, mobile data is a very promising data source as it does not only contain a large volume of information on the mobile user but also a large variety such as time, location, interests and contacts of the user. Therefore, our thesis claims that more useful insights can be extracted from mobile data for CRM issues using a top down approach.
Proposing a new framework, we define in part I the information on the user that can be retrieve from mobile data and they are organized in a framework called Mobile 3D (M3D) model. We suggest then a cartography of the research on mobile data and related ethical/legal issues in part II. Based on a literature review, the cartography highlights the level of interaction involved with the subject of interest and the focus level.
Finally, we investigate in depth and breadth the value of this data for CRM analytics in 3 empirical studies in part III. By depth, we mean to go beyond the analysis of a communication network by studying other mobile data such as url or mobility data, other featurisations of mobile data such as personality traits proxies and alternatives to this network. By breadth, we suggest to investigate this value not only to improve CRM issues for the telephone company which owns the mobile data but also for third party businesses outside the telephone industry. What is more, we propose a new algorithm called Essence Random Forest derived from the seminal Random Forest algorithm which is dedicated to handle redundancy of unstructured mobile data.
|Date of Award||29 Jan 2020|
|Sponsors||Université de Namur|
|Supervisor||Isabelle LINDEN (Supervisor), Philippe Baecke (Co-Supervisor), Jean-Yves Gnabo (President), Pietro ZIDDA (Jury), Martine George (Jury) & Dirk Van den Poel (Jury)|
Leveraging the value of cutomer mobile data in depth and breadth
Colot, C. (Author). 29 Jan 2020
Student thesis: Doc types › Doctor of Economics and Business Management