RetailCo — Discount Dependency and Net Revenue
How discount intensity shapes net revenue behavior across products, stores, and customer segments
SQL • Python • Tableau
Overview
This project examines how discount intensity affects net revenue performance in a retail context. Using transactional sales data across products, stores, and customer segments, the analysis investigates whether higher discount levels consistently lead to stronger revenue outcomes or if their impact varies depending on business conditions.
The goal is to assess discount dependency from a business perspective, identifying discount ranges that support revenue stability, highlighting diminishing returns from aggressive discounting, and uncovering differences in performance across stores and customer segments. The analysis combines data modeling, exploratory analysis, and visualization to support pricing and promotional decision making.
Findings are supported by an interactive dashboard that enables stakeholders to explore revenue behavior across discount levels and key business dimensions.
Business Problem
Retail businesses frequently use discounts to stimulate demand and increase sales volume. However, the relationship between discount intensity and net revenue is not always clear. While higher discounts may attract more customers, they can also compress margins, creating uncertainty about whether deeper discounts actually lead to better revenue outcomes.
This creates a core business question: to what extent does net revenue depend on discount levels, and is this dependency consistent across different business contexts? Without a clear understanding of this relationship, pricing and promotional strategies risk being guided by assumptions rather than evidence.
The challenge, therefore, is not simply to observe sales performance under discounts, but to understand whether increasing discount intensity meaningfully changes revenue behavior, or whether its impact varies depending on operational and customer related factors.
Key Metrics (KPIs)
• Net Revenue (primary KPI): revenue after discounts, used as the main outcome to evaluate how revenue performance behaves under different discount conditions and to assess potential dependency on discounted transactions
• Gross Revenue: pre-discount revenue baseline, used to contextualize overall performance and to distinguish revenue driven by underlying demand from revenue driven by price reductions
• Discount Rate (%): average discount intensity, derived from discount values, used to compare discount dependency profiles and to define discount buckets across segments
• Average Order Value (AOV): order level value indicator, used to assess revenue quality across different discount levels
Analytical Approach
• Integrated and validated transactional, customer, and store level data using SQL executed within a Python environment
• Built analytical aggregates to quantify discount distribution and net revenue behavior
• Analyzed net revenue performance across discount ranges, time periods, and business dimensions
• Consolidated insights through visual analysis and summary metrics that could support discount related insights
Selected Visuals
This visualization compares discounted and non discounted net revenue across the top performing products. The results show that discounted sales account for a significant share of net revenue even among high performing products, indicating that discounting is broadly applied rather than limited to underperforming items. This suggests that product performance alone does not fully explain revenue outcomes under discount driven strategies.
Store level results reinforce that discount dependency is widespread, but not identical in intensity. Some locations present more pronounced deviations, indicating local context matters (assortment, customer mix, operational patterns). This supports a targeted approach: reviewing discount behavior at the granular level is likely more actionable than applying store wide assumptions.
Net revenue behavior differs across customer segments. Contrary to common expectations, wholesale customers do not consistently generate higher net revenue compared to other customer types. This indicates that higher purchase volume does not automatically translate into greater revenue contribution and that discount dependency varies by customer behavior rather than customer category alone.
Over time, net revenue and discounted sales volume move together, reinforcing the structural nature of discount dependency. Peaks in revenue tend to amplify the same discount driven mechanism rather than introduce a fundamentally new pattern.
Interactive Dashboard
The interactive dashboard consolidates product, store, customer segment, and time based views into a unified analytical interface. It enables deeper exploration of discount distribution and net revenue behavior across multiple dimensions, complementing the static visuals presented above.
By interacting with the dashboard, users can dynamically filter discount ranges, customer types, and stores to identify patterns, exceptions, and segment specific behavior that are not immediately visible in static charts.
Key Findings
• Net revenue appears structurally dependent on discounts
• Meaningful variation tends to emerge at granular levels
• Wholesale customers do not systematically outperform other customer types in net revenue generation
• Temporal peaks tend to amplify the same discount-driven behavior rather than create new revenue dynamics
Tools & Technologies
• SQL
• Python (DuckDB, Pandas, NumPy, Matplotlib)
• Data Visualization: Tableau
Analytical Scope and Limitations
This analysis is exploratory and descriptive in nature, focusing on identifying patterns and relationships between discount levels and net revenue within the observed dataset. While the results highlight meaningful revenue behaviors across different discount ranges and business segments, they do not define optimal pricing or discount strategies.The findings should be interpreted as analytical signals that help frame pricing related discussions rather than prescriptive recommendations.
Full Project Repository
The complete project, including data preparation, analytical queries, exploratory analysis, and documentation, is available in the GitHub repository. The repository contains the full Jupyter Notebook and detailed explanations supporting each step of the analytical workflow.