Customer Analytics in FMCG Industry

by Sooyeon Won

Keywords

• Marketing Mix
• STP framework
• Customer Analytics
• Segmentation and Clustering
• Dimensionality Reduction with PCA
• Data Visualisations
• Purchase Analytics
• Price Elasticity
• Modeling Purchase Incidence
• Modeling Brand Choice
• Modeling Purchase Quantity

Introduction

This project is about customer analytics in Fast-Moving-Consumer-Goods (FMCG) Industry. The project is consisted of Customer Analytics and Purchased Analytics. It is motivated by Customer Analytics Program in Udemy.

• Customer Analytics: The first part of analysis focuses on how to perform customer segmentation. It involves the application of hierarchical and flat clustering techniques for dividing customers into groups. It also features applying the Principal Components Analysis (PCA) to reduce the dimensionality of the problem, as well as combining PCA and K-means for customer segmentation.

• Purchase Analytics: The second part of analysis explores both the descriptive and predictive analysis of the purchase behaviour of customers, including models for purchase incidence, brand choice, and purchase quantity.

Summary of Findings

• Segmentation: It turns out that it is the most appropriate group all datapoints into four segments. Each segment has different characters : 'Well-Off', 'Fewer-Opportunities', 'Standard', 'Career_Focused'. "Fewer-Opportunities" segment is the largest. Almost 40 percent of customers belong to this segment.
• Brand Preference: Each segment has also unequal brand preference. Fewer-Opportunities segment shows an extremely strong preference for brand 2. Ca. 63 percent of the career focus segment buys brand 5 which is the most expensive brand. Well-Off segment enjoys one of the most luxurious brands but not the most expensive one. Brand 4. Standard segment is the most heterogeneous segment.
• Revenue: Brand 5 brought the most revenue. On the other hand Brand 3 has the lowest revenue, although it is not the cheapest. "Career-Focused" brings the most revenue.
• Price Elasticity: The higher the price of our product becomes the less likely it will be for people to want to buy it. "Fewer Opportunities" segment is the most price sensitive compared to the average as well as other segments.
• Promotion: People are more willing to buy products at promotional prices and they are a little bit more elastic when there is a promotion.

Referenecs

Price Elasticity Customer Analytics