Key takeaways
- How a telecom churn model illustrates the shift from a black-box prediction to an explainable one that marketing and business teams can trust
- Why marketing is moving from scheduled, reactive campaigns toward real-time, data-driven customer interactions
- How machine learning interpretability applies to marketing attribution, showing which channels and spend levels actually drive return on investment
- How attribution insights inform budget allocation decisions across channels such as television and online advertising
- How demand and budget forecasting helped retail and supply chain teams adapt to shifting buying patterns
- Why the people and process around a model, not just the algorithm, determine whether a marketing AI project succeeds once it reaches production
Overview
In this session, H2O.ai's Bart Bakker looks at how marketing teams are putting machine learning to work in day-to-day strategy and budget decisions. The recording uses a telecom churn model as a working example of moving from a black-box prediction to one a business team can trust. It covers marketing attribution, using interpretability techniques to see which channels and spend actually move return on investment, and how those findings feed budget allocation across TV and online. It closes on recommendation engines and real-time personalization as the CMO's next step.