AI-Assisted Media Planning
summary
A deployed booking tool automates linear and streaming TV buys and embeds machine learning to predict efficiency and availability. Cost-per-visitor is modeled as a predicted CPV index (inverse CPV ×100) using a lightGBM model with ~50–100 features; simpler aggregate features dominate. Clearance uses a linear regression baseline that hits 85–90% of actual clearance and a lightGBM next model that improves accuracy by ~10–15%. Roadmap includes network look-alikes and automated self-optimizing "self-driving" plans.
AI-written summary · about
chapters (9)
- 0:00Introduction to AI media planning
- 1:06Tatari’s TV ad buying context
- 4:17What is media planning
- 6:41Traditional vs AI-driven TV buying
- 9:05Data science in media tools
- 10:06AI algorithms for CPV prediction
- 13:19Clearance estimation models
- 16:53Future of self-driving ad plans
- 18:58Key takeaways and Q&A
related talks
30mAI in Retail - Advancing Promo Demand Forecasting
29mAI and ML for Computational Advertising
29mAI & ML in tailored software for decision-making purposes
25mHow we bring Performance Marketing to linear TV by using machine learning
27mGenerating targeted AI revenue at mass scales
41mUnlocking Profit: How AI-Driven Budget Optimization Elevates Marketing Performance
24mAI-Powered Transformation in Fashion E-Commerce
14mAI-Video Ads – Built 90% Faster & 80% Cheaper