Modeling Marketing Mix using PyMC3

Experimenting with priors, data normalization, and comparing Bayesian modeling with Robyn, Facebook’s open-source MMM package

In this article, I apply a Bayesian approach to the marketing problem of estimating the impact of advertising spend in different media channels on revenue. I cover several aspects of Bayesian modeling which should be important to the MMM practitioner:

• Normalization of dependent and independent variables and selection of priors
• The influence of normalization of the response variable on …

bayesian inference marketing marketing-mix-modeling modeling pymc3