GoNetspeed’s profitability depends on take-up-rate — the pace at which new customers sign up after build out. RDSC applied advance data science models to reduce risk by forecasting regions that would produce the take-up-rates to warrant investment in infrastructure.
Partner
GoNetspeed
Industry
Fiber-to-the Home
Objective
Locate optimal location for capital expansion
Approach
Machine learning
Statistical Modeling
Data Augmentation
The Opportunity
Fiber-to-the-Home (FTTH) provider GoNetspeed sought to expand its business in the US Northeast and beyond. Optical fiber installation is capital intensive and requires significant investment of both time and money towards right-of-way procurement and infrastructure construction before any cable is strung. If GoNetspeed could forecast the regions with the highest take-up-rate – the pace at which new customers sign up in a build-out area — they could substantially decrease the risk of its capital investment.
The Challenge
GoNetspeed experienced varying take-up rates within its current network boundaries. RDSC’s task was to determine why rates were higher in certain areas than others. Then, applying advanced modeling, project the findings to predict optimal areas for expansion.
“RDSC provided data driven insights that helped us to identify opportunities for expansion and minimize the risk of capital investment in our infrastructure build outs.”
The Solution
RDSC analyzed current customer data that including the various stages of the sign-up process (e.g., installed versus expressed interest), along with demographic and socioeconomic characteristics from the U.S. Census Bureau, and ArcGIS Tapestry ™ data, a proprietary dataset that describes regional markets according to lifestyles dominating those regions (e.g., empty nesters). Robust data exploration underpinned the process and ensured critical details about the business were included in the analysis. Statistical and machine learning modeling techniques were applied to explain take-up rates in the current network and to predict where high uptake rates were probable beyond the current network in the northeastern US. A list of metropolitan areas ranked by the take-up rate forecasted was provided.
The Deliverable
RDSC provided powerful decision support through a geographical information system that enabled GoNetSpeed to explore the findings firsthand. Analysts were able to view maps of forecasts and associated socioeconomic factors which led the model to predict take-up rates. Regions were color-coded according to projected growth rate, enabling analysts at GoNetspeed to quickly identify regions of interest.
