Summary

In this chapter, we dove deeper into what a media mix model is and what it is used for. We discussed what data we should gather and how to transform it. We also discussed different adstock models and how to account for media saturation. Finally, we went through an example of how to fit a media mix model using PyMC Marketing with synthetic data, to understand how we could recover the original parameters.

You now know how to discuss and evaluate a media mix model, understanding its strengths and limitations. You can extract the correct data needed to operationalize one, ensuring reliable inputs. You can also implement a simple media mix model by using the PyMC library and applying Bayesian methods, as well as calibrate the model to improve its accuracy and predictive quality in marketing analysis.

In the next chapter, we will go through A/B testing, and how to use it to measure the impact of media on sales.