Data Science

Behavioral segmentation involves grouping customers into segments based on behavioral attributes
This is different from other segmentation approaches (like RFM)
- Since, RFM doesn't include behavioral attributes
- Instead, RFM only includes engagement and monetary attributes
In other words, behavioral segmentation aims to identify the behavioral attributes that cause monetary attributes (like RFM)

Usually, this involves deliberately excluding monetary and engagement attributes from clustering
- After clustering, these attributes are included in customer profiles
- By doing this, we ensure segments are generated based on the behavioral attributes
- Thus, we're essentially determining the behavioral causes of these monetary and engagement attributes (rather than vice versa)
Also, this allows us to include behavioral attributes for both clustering and profiling purposes more easily
- These attributes are important for helping us understand the drivers behind customer behavior
- E.g. why one segment churns more than another segment

Segments can provide an important, targeted signal in propensity models that may be missed otherwise
- These signals are carefully targeted and discovered using clustering methods
- Whereas, these signals may be missed by just throwing a bunch of behavioral features into a classification model
Creating more targeted classification or propensity models
- Classification and propensity models created for an entire customer population usually have limited accuracy
- Since, propensities can be determined by many different factors
- For example, a customer in one segment may churn due to low product quality
- Whereas, a different customer in another segment may churn due to high prices
- Consequently, the model repository can maintain specialized models for different combinations of:
  - Business objective
  - Brands
  - Product classification
  - Customer segments

Propensity Modeling

Overview of Segmentation