Objective of the project was to predict the likelihood of a customer buying a specific brand for a particular email campaign.
It consists of a combination of 3 models.
- M1 - Propensity towards purchasing products in general
- M2 - Propensity towards purchasing from a specific brand
- M3 - Propensity to make purchases in the following month
The data that was fed into the model included customer level data assets. The data assets described the spending behaviours, socio demographic information, product information and campaign-level information. Feature engineering was done prior to building the models.
AWS Sagemaker/Python were leveraged for the feature engineering as well as model building.