Predicting Customer Churn for DTH: Building Churn Score Card for DTH

Ankit Goenka (), Chandan Chintu () and Gyanendra Singh ()
Additional contact information
Ankit Goenka: Experian Credit Information Company of India Private Limited
Chandan Chintu: Experian Credit Information Company of India Private Limited
Gyanendra Singh: Experian Credit Information Company of India Private Limited

A chapter in Advances in Analytics and Applications, 2019, pp 85-104 from Springer

Abstract: Abstract Customer churn is when the customers either switch from their incumbent service provider to its competitor or stop using the service altogether. Per existing research, it costs five to six times more to acquire customers than to retain existing customers, and this emphasizes the importance of managing churn by organization. In this paper, we build Churn prediction model for one of India’s largest Direct to Home (DTH) operator, for its customer base. We use data provided by the DTH operator to build the model. Given the varied base of customers, the data was segmented in smaller homogenous chunks, with similar profile and behaviour. To achieve this, we conducted segmentwise analysis to determine the factors that affect churn differently in different segments. This indicated the need to have different models. Analysis showed that age on the network was the key segmentation driver and hence separate models were built for each segment. We further applied various data mining techniques such as logistic regression, random forest and gradient boosting method to build the model for each segment. Gradient boosting method outperformed both logistic regression and random forest on measures of AUC and Gini. The proposed model correctly classifies churn customers between 76 and 78% depending on the segment. The primary drivers of churn across all the segments for DTH are the age of the customer, type of package subscribed, longest delinquency, maximum amount recharged and maximum valid days for which customer has recharged their set up box. The paper also shows that customer experience while interacting with operator and quality of device are equally important attributes.

Keywords: DTH; Churn; Machine learning (search for similar items in EconPapers)
Date: 2019
References: Add references at CitEc
Citations:

There are no downloads for this item, see the EconPapers FAQ for hints about obtaining it.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:spr:prbchp:978-981-13-1208-3_9

Ordering information: This item can be ordered from
http://www.springer.com/9789811312083

DOI: 10.1007/978-981-13-1208-3_9

Access Statistics for this chapter

More chapters in Springer Proceedings in Business and Economics from Springer
Bibliographic data for series maintained by Sonal Shukla () and Springer Nature Abstracting and Indexing ().