Capstone Projects

Predicting Broadband Customer Take Rates in New Markets

Program: Data Science Master's
Host Company: Telephone and Data Systems
Location: Madison, Wisconsin (onsite)
Student: Mark Hoferle

The telecommunication industry has been evolving for several years from selling voice services in a regulated environment to becoming a broadband service provider in a deregulated market. Traditional phone companies now offer internet services in new geographic markets to a new base of customers and face new competitors. As telecommunication companies develop business cases and financial modes when deciding to enter a new market, it’s important to accurately forecast the number of customers that will purchase their services. This paper outlines the steps taken to develop two different classification models to predict whether a given household (service address) in a new market will become a customer. Various address attributes (customer demographics, existence and type of competitor, etc.) were incorporated into the models to determine which combination of attributes is the best predictor. The paper outlines the observations and conclusions reached from this effort and describes next steps that could be taken to extend this analysis.