Customer Behavioral Segmentation
Behavioral segmentation of retail customers aimed at improving response to price promotions
Problem Statement
A US-based apparel retail company wanted to identify opportunities to improve responsiveness to price promotions. Their existing customer segmentation, based solely on demographics, failed to capture purchase behaviors and unmet needs that drove purchase decisions. To maximize promotional ROI, a new behavioral segmentation scheme that allows for targeted and personalized price promotions was required.
Challenges
- Siloed data sources requiring extensive data gathering and transformation effort
- Designing analytical approach to overlay responsiveness towards specific price promotions on top of behavioral segments
Solution
Summary
I, along with a team, developed an unsupervised machine learning solution to identify behavioral clusters in the customer data. The segments were profiled to understand intra-cluster commonalities and responsiveness to different types of price promotions was evaluated for each segment.
Approach
- Hypothesizing key behavioral metrics such as spend, frequency of purchases, basket size, basket components, average price of products, etc.
- Feature selection and variable transformations
- k-means clustering to identify customer segments
- Descriptive analysis on select behavioral metrics to profile segments
- Evaluating the sales lift of each segment during each promotion period in the preceding one-year time horizon
Data Used
- Point of sales data
- Customer demographics
- Product attributes
- Store attributes
- Socio-economic indicators
Key Takeaway
- Two types of high value customers were identified - High Volume and High Frequency customers. This was a novel actionable insight to the marketing team.
Deliverable
- Segmentation and analysis report with key findings and executive summary
- Mapping of segments with existing customer base
K-Means Clustering - Illustration
Promotion Response Sales Lift - Illustration
Have a similar challenge?
Let's discuss how we can develop solutions for your specific use case.