• DOI:10.1007/S11042-021-10897-X
• Corpus ID: 236256590

Capturing the real customer experience based on the parameters in the call detail records

@article{Khan2021CapturingTR,
  title={Capturing the real customer experience based on the parameters in the call detail records},
  author={Nusratullah Khan and Muhammad Usman Akram and Asadullah Shah and Norah Saleh Alghamdi and Shoab Ahmed Khan},
  journal={Multim. Tools Appl.},
  year={2021},
  volume={80},
  pages={28439-28461}
}

• Nusratullah Khan, M. Akram, +2 authors S. Khan
• Published 4 June 2021
• Business
• Multim. Tools Appl.

Customer Experience-Management (CEM) has emerged as a key differentiator in recent fierce market of telecommunication. A positive customer experience leads to increased loyalty, lower churn rate, more recommendations and optimistic word of mouth. In this era of technology, digital data has become an asset of any business. This paper proposes a technique that calculates the Customer Experience Management Index (CEMI) of subscribers of cellular network services providers by using their Call Detail Records (CDRs). In the first phase, CDR Dataset of subscribers, using the same cellular network operators, is collected. In the second phase, a close-ended telephonic survey is prepared and conducted on these subscribers of targeted valued community or social network. Six attributes of telecommunication service are selected i.e. network coverage, voice call quality, drop call rate, Short Message Service delivery, internet service and call setup duration. The subscribers are requested to grade each attribute as per their experience while using that service and to rate the service collectively in order to identify the overall experience. Genetic Algorithm (GA) is applied to optimize the weights for each attribute to eventually formulate a mathematical model for CEMI calculation. The proposed technique uses the difference between weighted attributes based on calculated CEMI and actual CEMI provided during survey process as cost function for GA which updates weights in such a way to have minimal value of this cost function. Same attributes are selected from CDRs and they are graded on the same scale which is used in survey based on the events and flags presented in the CDRs. While concluding the results, the obtained optimized weights are applied to the technical attributes of the CDRs data set and the CEMI is calculated. The proposed technique calculates the CEMI in real-time with an accuracy of 84.5 % and error of 0.248.