E-Commerce Insights Dashboard & Python Machine Learning Predictions
Know Who’s Leaving Before They Do
Retention starts with recognition. I built this project to help businesses not just track sales, but understand customers — and take action before it’s too late.
Most teams rely on lagging reports or guesswork when it comes to churn. This system blends interactive dashboards and AI-powered churn modeling to turn raw data into early warnings.
You’ll spot which customers are drifting away, what’s driving loyalty, and how to fine-tune your outreach — all in one clear view. It’s not just reporting; it’s predictive insight.
Because when you know who’s slipping through the cracks, you can stop the leaks — and grow smarter, not just bigger.
Health & Fitness - SMB Supplements Company (50 employees)
Industry: Health & Fitness
Annual Revenue: $500k
Tools:
Power BI - Data Visualization & Report Distribution
Python (Pandas, Scikit-learn, Matplotlib) - Data Manipulation, Forecasting, & Visualization
BI Maturity: Low – Data spread across systems, no centralized reporting
Problem:
This supplement company had growing sales but lacked clarity on what was really driving performance. Customer behavior was hard to interpret, ad ROI wasn’t tracked consistently, and they had no way of identifying churn risk. Most decisions were reactive, based on siloed spreadsheets or gut instinct — not data. As the business scaled, the lack of centralized, actionable insights made it difficult to optimize marketing, retain customers, or measure true profitability.
Project Goals:
Centralize sales, customer, and ad data into one interactive dashboard
Visualize key metrics like AOV, ROAS, CAC, and LTV
Segment customers by lifecycle stage: New, Active, Churned
Predict churn risk using behavioral data and machine learning
Enable leadership to spot trends and take proactive action faster
Interactive Power BI dashboard tracking revenue, customer behavior, ad performance, and churn risk — built to help e-commerce teams make faster, smarter decisions.
Churn probability scores generated by the model — each row represents a customer, with the final column showing their predicted likelihood to churn.
What Happened?
Clarity From Complexity
The dashboard unified sales, marketing, and customer data — turning scattered reports into a centralized source of truth. Teams finally had real-time insight into how the business was performing.
Smarter, Data-Led Conversations
Marketing and leadership began using the dashboard to guide weekly reviews, uncovering which channels, products, and customer segments were really driving revenue and retention.
Churn Risk Exposed
The churn prediction model identified high-value customers slipping away — giving the team a chance to act before revenue disappeared. This led to targeted win-back campaigns and improved customer lifetime value.
Decisions With Confidence
With AOV, CAC, ROAS, and LTV all in view, the company could finally tie spend to impact. They stopped reacting to lagging indicators — and started planning ahead with precision.
Related Blog Posts
-
![]()
Blog #1: Businesses are great…ha…maybe not always. But let me try to explain why I love them.
