This session will discuss Wise’s approach to Bayesian Marketing Mix Models & Marketing Attribution Triangulation, with a particular focus on using multiple sources of data, stakeholder knowledge and experiments to help calibrate these models.

It will be a mixture of high level and technical discussion. The methodology should be interpretable to audiences who are interested in kicking off Marketing Mix Model Projects at their companies, with code snippets & examples particularly interesting to Data Scientists and Analysts wanting to take their MMM’s or similar Econometric Bayesian Regression Models to the next level.

There will be a brief intro to Marketing Mix Models & Bayesian Methodologies as a concept to those unfamiliar. We will cover common Measurement Pitfalls such as holidays, small channels, modelling trend/seasonality in rapidly growing markets, insignificant geo-experiment test results, limited data, as well as their solutions.

Technical Level: Advanced technical/mathematical