Olist Retention Analysis
Project Walkthrough
Slide 1 – Title
Slide 2 – Business Problem
Slide 3 – About Olist
Slide 4 – Retention Crisis
Slide 5 – Dataset Scope
Slide 6 – Data Engineering
Slide 7 – Power BI: Business Performance
Slide 8 – RFM Methodology
Slide 9 – Power BI: Customer Analytics
Slide 10 – R$4.6M Opportunity
Slide 11 – Power BI: Products & Geography
Slide 12 – Delivery Delays
Slide 13 – A/B Test Design
Slide 14 – Power BI: A/B Test Results
Slide 15 – Business Impact
Slide 16 – Recommendations
Slide 17 – Key Takeaway
Slide 18 – Thank You
1 / 18

Olist had a catastrophic retention problem: 97% of customers never returned for a second purchase, against an industry benchmark of 25–30%. A structural revenue ceiling that no amount of new acquisition spending could fix.

Starting from 9 raw CSV files, I built a medallion-architecture pipeline, scored 93,358 customers via RFM segmentation, and identified a R$4.6M revenue opportunity. A simulated A/B test validated that a targeted 10% coupon delivers a +38.6% lift in repeat purchases (p<0.001) with 186% ROI and payback in Month 1.

Tech Stack
Python Pandas SQL Power BI DAX RFM Segmentation A/B Testing Medallion Architecture
View on GitHub Read Article