Challenge
1 —
A cosmetics giant in an environment where promotion is key and competition is increasingly increasing promotions
2 —
A willingness to get out of this dependence on promotions by better distributing the discount to customers according to their expectations
3 —
A multi-team project involving Data Science, Marketing and Campaign Management
The Key Questions
1 —
Which promotions to send to which customers?
2 —
How to measure the impact of personalized promotions?
3 —
What optimizations can be made to the allocation of offers?
4 —
How to industrialize the process?
Approach
1 —
Develop the model that predicts activation and margin expectation for each type of promotion
2 —
Set up a training methodology for the model and testing
3 —
Measure performance during each test phase
4 —
Identify potential recommendations and optimizations and support the transition to industrialization
Results
1 —
A 1.5 pts decrease in discount rate and a 3% increase in margin
