Poster
Client Context
Cox Communications is the 3rd largest cable provider in the US. They serve 3 customer groups: residential, small business, and enterprise. Cox’s pricing team handles setting product prices on a monthly basis. Our focus is on pricing internet products to enterprise customers.
Project Objective
Redesign Cox Communications Inc.’s enterprise pricing process to create an automated platform that produces more competitive bids, while incorporating more data and new data analysis techniques.
Design Strategy
The project was divided into two phases. Phase 1 sought to replicate, automate, and improve upon Cox’s current pricing process through the creation of our Pricing User Interface (PUI). Phase 2 sought to return a new recommend price based on probability of winning (Bid Confidence Model) and evaluate this price change (Trend dashboard). Identifying a bid’s probability of winning allowed us to maximize the expected revenue per bid and use this as a recommended price.
Deliverables
(1) Pricing User Interface – Python GUI that completely automates client’s pricing process eliminating the technical barrier, while adding outlier handling and improved regression techniques.
(2) Bid Confidence Model - a Random Forest regression model that predicts the probability of a bid winning given price, region, internet speed, and other features. It returns the price that maximizes the expected revenue curve.
(3) Trend dashboard - a Tableau dashboard that evaluates market changes such as changes in competitor pricing year over year, Cox's win rate versus other competitors, and more.
Value and Impact
Estimated 750 hours saved per year. 28% average revenue increase in high-data regions based on BCM. 400X process speed improvement over current process with PUI.
Project Information
Student Team
Will Allen, Anna Blackwell, Will Enger, Brett Graybill, Sam Knobbe, Isaac Zelcer
Faculty Advisor
Faculty Evaluator
