E-Commerce Sales Analysis

I've been building an e-commerce analytics report in Power BI covering 48,000 transactions, 4,000 customers and 101 software products (Apr 2024 – Oct 2025). It now has five pages: Executive Summary, Customer Loyalty, Revenue Drivers, Customer Health & Retention, and Refunds & Promotions.

Some of the insights the data surfaced:

Annual plans drive the business. They're 49% of orders but 88% of revenue, and each annual order is worth 10× a monthly one.
Big deals matter most. Orders of 10+ seats are 29% of volume but 70% of revenue.
Revenue is concentrated. 38 of 101 products generate 80% of revenue.
Retention is strong. About 85% of customers keep buying each quarter, but new customer acquisition almost stopped in 2025.
There's a clear win-back list. 492 high-value customers have gone quiet, and together they hold $5.9M of lifetime revenue.
Over half of refunds are preventable. 56% of refunded revenue came from billing errors and duplicate orders.
Promo codes are leaking. Welcome codes were reused on ~5,000 repeat orders ($319K in discounts), and 89% of Black Friday codes were redeemed outside Nov–Dec.

What I used:

• DAX for RFM segmentation, cohort retention, like-for-like YoY and Pareto analysis
• Deneb (Vega-Lite) for custom visuals: Pareto chart, cohort heatmap, dumbbell chart, refund hotspot matrix
• HTML Content visuals for KPI tiles with sparklines and dynamic insight cards
• SVG measures to put data bars and bullet charts inside tables
• A consistent custom theme, background and typography across every page

Lesson learned: auditing the model before adding visuals paid off. It caught a discount measure totaling $61.9B instead of $1.26M, discount rates above 100%, and tax counted twice in gross revenue.

Feedback welcome!

#PowerBI #DataAnalytics #DAX #Deneb #DataVisualization #Ecommerce #BusinessIntelligence