Forecasting Media Customer Behavior
The Challenge
A media company wanted to introduce new and enhanced audience-based products to execute advertisement deals and enable clients to optimize how their ad dollars work. To do this, they needed to forecast the demand for multiple advertising choices by using data and insights from different sources. However, using traditional methods, they faced challenges such as data inconsistency, uncertainty, and complexity.
The Solution
We designed a dynamic forecasting process that uses data fusion and hurdle predictive models to estimate customer behavior and preferences. Data fusion is the process of integrating multiple data sources to produce more consistent, accurate, and useful information. Our solution helps the Spot Scheduler in scheduling advertising spots across different networks and programs.
The Results
Our solution improved forecasting accuracy by 13% and helped the company to expand its advertising service to five different countries across Latin America.