Predicting advertising returns with AI
Use Case: A BLOCKS system for making predictions that any employee can use.
One of the pillars of Grand Vision’s “Business Production” business is their “Mail-Order Production” service, in which they help plan and create advertisements and mail-order businesses. As part of this service, Grand Vision employees called “Planners” would make predictions for advertising returns and plan their time-slot purchases accordingly. However, training Planners who could accurately predict returns and select good time slots was a big investment, and even the most experienced Planners could sometimes make mistakes in their predictions.
Grand Vision used their past results data, as well as accumulated data from their call center (which handles the actual calls and orders from customers who view the advertisements), to train a regression model with the MAGELLAN BLOCKS Model Generator. While their earlier predictions has about 80% accuracy, the new model could predict returns with over 95% accuracy.
Since BLOCKS simplified the process for making predictions, Grand Vision is now looking to reorganize so that any employee—not just the highly-trained Planners—will be able to make predictions and provide highly accurate proposals. They have now used their know-how from this experience to create a new service for predicting advertising returns they call a GAM Report (Grandvision AI Marketing Report).
- Model GeneratorRegression
- Flow Designer