Kollex – Customer Reactivation Engine
How I built a predictive segmentation engine that identified high-value dormant clients and empowered sales teams to recover lost revenue.
ABOUT THE PROJECT
Overview
Transforming raw transaction data into a live "Health Check" for the customer base. I replaced intuition-based sales calls with data-driven targeted outreach.
The Challenge
The sales team was "flying blind." They had thousands of inactive customers but no way to know which ones were worth chasing. They wasted hours calling low-value leads while high-volume accounts quietly churned to competitors without anyone noticing.
The Solution
- RFM Segmentation (Python):: I developed an algorithm based on Recency, Frequency, and Monetary value to automatically categorize 10,000+ restaurants into segments like "Champions," "At-Risk," and "Dormant."
- Churn Prediction Model:: I analyzed order patterns to flag customers who were deviating from their usual buying cycle (e.g., "Missed their usual Monday order"), triggering an alert for the sales team.
- Automated Lead Scoring:: Instead of calling random leads, sales agents received a daily "Hit List" of high-value customers who were most likely to re-order.
- CRM Integration:: Synced these insights directly into HubSpot, allowing agents to see a customer's "Health Score" and recommended products before making the call.
I moved beyond static reporting and built a Predictive Sales Engine:
Let's talk about your data
Every engagement starts with a short call. I'll look at your setup and tell you exactly what I'd build, what it costs, and how long it takes.
Free 30-min data review