1,000+
4
26
>90%
score in metrics
Key challenges
A leading beverage company struggled to decode consumer behavior across categories, with siloed analytics limiting insight into awareness, perception, and consumption. To move from assumptions to evidence, we applied machine learning to survey data from 1,000+ consumers, classifying segments, uncovering key drivers, and empowering teams to make targeted, data-driven marketing and product decisions.
Manual handling limiting scalability and insight generation
Difficulty linking consumer behavior to marketing strategy
Data complexity and non-actionable variables
Unclear understanding of consumer motivation
No predictive insight into future trends
The solution
Data engineering and modeling
Built 4 category models
Grouped and refined features
Applied PCA to 26 key variables
Insight generation
Linked behaviors to levers
Used SHAP and dependency analysis
Revealed influence and quantified feature contribution
Data preparation
Cleaned 1000+ surveys
Grouped features and filtered
Modeling
Built category models
Optimized via grid search
Activation
Used explainable AI
Shared insights enterprise-wide
Marketing and digital Initiatives
Optimized digital and sponsorship campaigns
Enhanced consumer reach and engagement
Product strategy
Guided product improvements based on quality perception
Aligned packaging and pricing with consumer preferences
