Protecting Market Share with Predictive Analytics and Smart Clustering for an FMCG Giant

Introduction

A well-established Fast-Moving Consumer Goods (FMCG) company was facing a major challenge: a new market entrant which was a competitor in the same category, was gaining traction, taking up valuable shelf space, and threatening their customer acquisition cost (CAC).

To counter this threat, the client needed a data-driven approach to customer journey optimization—one that would enable them to:
✅ Identify which stores were at risk of switching to the competitor Brand.
✅ Allocate promotional budgets efficiently, ensuring the best possible Customer Experience ROI.
✅ Equip their sales teams with a simple, actionable tool to respond in real time.

Numr CXM designed a predictive analytics solution that leveraged CX data analytics and advanced statistical modeling to pinpoint at-risk stores and guide strategic interventions.

The Challenge: Identifying Vulnerable Stores Using Limited Data

The client faced three core challenges:

1. Determining Store Vulnerability Without Direct Customer Insights

The only available data sources were stocking and sales reports, making it difficult to predict which stores might switch to the competitor brand. CX strategy needed to be data-driven without requiring direct surveys or additional research investments.

2. Avoiding Wasted Promotional Spend

Running promotions across all stores was not a feasible option—there was a cost associated with each intervention. The goal was to ensure only vulnerable stores received targeted sales promotions, maximizing the ROI of CX efforts.

3. Providing Sales Teams with a Simple, Actionable Solution

The company’s sales representatives needed an easy way to identify at-risk stores without relying on complex data models. A straightforward, on-the-ground assessment method was necessary for effective implementation.

‍

The Solution: A Data-Driven CX Strategy

To solve this challenge, Numr CXM developed a multi-stage, predictive analytics approach that would:
✔ Cluster stores based on stocking patterns, identifying common characteristics of high-risk stores.
✔ Determine which brands were most frequently associated with the Competitor Brand, helping sales reps identify vulnerable locations.
✔ Predict future store conversions, enabling preemptive action through targeted promotions.

1. Clustering Stores to Identify At-Risk Locations

The first step was to analyze stocking data using K-Mode Clustering, grouping stores based on the brands they carried.

• This revealed three distinct store clusters, explaining 70% of the variance in stocking patterns.
  
  
• However, brand overlap between clusters made it difficult to clearly differentiate high-risk stores.
  
  

To refine the model, we ran a Correspondence Analysis using sales volume data instead of just availability. This uncovered:
✅ Stronger brand associations, highlighting which stores were more likely to add the competitor Brand.
✅ Key risk indicators, making it easier to classify stores using a simple brand check.

2. Predicting Which Stores Would Start Stocking the Competitor Brand

To predict future risk, Numr CXM applied Predictive Analytics:

1. Analyzed stores already stocking the Competitor Brand and tracked their historical stocking patterns.
   
   
2. Identified an inflection point—the moment when a store switched to stocking the competitor Brand.
   
   
3. Created a Predictive Model that assigned a risk score to each remaining store.
   
   

This model flagged 170 stores in the city that were not yet stocking the new competitor brand but were highly likely to do so soon.

3. Simplifying the Model for Sales Teams

The final step was to make the insights actionable for on-the-ground sales reps.

• Numr CXM designed a simplified version of the model, allowing sales reps to identify at-risk stores simply by checking stocked brands.
  
  
• This eliminated the need for complex data interpretation, ensuring real-time decision-making without additional tools.
  
  

The Result: Improved Market Position and Optimized CX Investments

🚀 70% accuracy in predicting vulnerable stores, allowing the company to take proactive action before losing market share.
💰 Maximized Customer Experience ROI by ensuring promotional budgets were used efficiently, only in high-risk locations.
🔍 Improved customer retention and customer lifetime value (CLTV) by keeping the client's products in key retail spaces.
📈 Strengthened CX strategy, proving the value of CX data analytics in competitive market environments.
🛒 Simplified in-store decision-making, giving sales teams a practical tool for customer journey optimization.

By leveraging predictive analytics and customer experience metrics, the client transformed its market defense strategy, proving that CX-driven data intelligence can significantly improve business outcomes.