Dataset: Kaggle Global Superstore (2021–2023) · 10,479 clean orders · 3 regions, 3 categories
Average discount tripled from 9.9% → 30.8% in 2023. At >30% discounts, 96–100% of orders generate a loss. This single factor accounts for ~70% of the revenue and margin collapse.
Central region applied 42.5% avg discounts vs the company's 9.9% baseline. Revenue fell from $2.05M → $0.81M. Likely a rogue promotional campaign or mis-configured pricing rules.
Furniture was already at −7.9% margin in 2021 before discounts. Tables alone lost $243K in 2023. Deep discounting on inherently low-margin products amplified losses catastrophically.
Revenue peaks in Q4 (Nov) and troughs in Q1 (Jan) across all years — consistent seasonality. However, 2023's drop is uniform across all months, ruling out seasonality as a cause. The issue is structural, not temporal.
2022 showed healthy +17.5% growth. The pivot to aggressive discounting in 2023 was clearly a strategic decision — possibly to boost order volume — that backfired severely. Order count only fell 17.8% while revenue fell 49%, confirming that discount depth, not lost customers, caused the collapse.
Source: Kaggle Global Superstore | 10,560 raw rows | 13 columns | 474 missing values | 51 duplicates | 30 negative sales entries
Used df.drop_duplicates() to remove 51 exact-match rows — caused by upstream ETL reprocessing errors. Keeps only the first occurrence.
30 rows had negative Sales values, indicating data-entry errors (not returns — returns would appear in a separate refunds table). Filtered with df[df['Sales'] > 0].
Ship Mode, Segment, and Region each had ~158 nulls (1.5% of data). Since they're categorical with dominant modes, imputed with mode value. Alternative: flag as "Unknown" if reporting requires it.
105 rows had profit values 25× their actual values — clearly erroneous multipliers. Capped at the 99th percentile ($1,783) using df['Profit'].clip(upper=p99). Winsorization preferred over removal to preserve order records.
Parsed Order Date and Ship Date from strings. Derived: Year, Month, Quarter, Month_Name, and Lead_Days (ship lag). These enable all time-series and seasonal analyses.
10,479 rows · 18 columns · 0 nulls · 0 duplicates · 0 negative sales · profits capped · dates parsed. Ready for analysis and dashboard loading.
Revenue fell −49% in 2023, driven by an uncontrolled discount expansion that tripled average discounts from 9.9% → 30.8%. This, combined with structurally unprofitable Furniture products and a complete breakdown in the Central region's pricing discipline, turned a healthy +17.5% growth trajectory into a −$726K loss year.
Set system-enforced maximum discounts: Technology ≤ 15%, Office Supplies ≤ 20%, Furniture ≤ 10% (already unprofitable at baseline). All discounts above tier 2 should require VP of Sales approval.
🔴 High ImpactCentral's 42.5% avg discount is 4× the baseline. Immediately freeze all promotional pricing in Central, conduct a pricing audit, and retrain the regional sales team on margin-based selling. Investigate whether pricing rules were mis-configured in the order management system.
🔴 High ImpactTables had a −34.6% margin even before 2023's discounting. The product either needs to be repriced 40%+ higher, bundled with higher-margin items, or discontinued. At current pricing, every Tables sale destroys value.
🔴 High ImpactSales reps likely received revenue-based commissions, incentivizing them to close deals with deep discounts. Restructure comp plans to reward gross profit dollars. A $10K deal at 5% margin is worse than a $6K deal at 25% margin.
🟡 Medium ImpactImplement daily monitoring of avg discount rates by region and category. Trigger automated alerts when a region's avg discount exceeds the baseline by >5pp. This case study's damage went undetected for 12 months — a dashboard would have caught it in week 4.
🟡 Medium ImpactArt (+9.6% margin) and Labels (+4.7%) remained profitable in 2023 despite the discount push. Accessories nearly broke even. Expand assortment and marketing investment in these categories while reducing exposure to Computers and Tables which are margin-negative even at low discounts.
🟢 StrategicReturning to 2022 discount levels (9.9% avg) while maintaining 2023's order volume (3,194 orders) would recover approximately $4.8M in revenue and restore margins to +1.4% to +3% — yielding estimated profit recovery of $700K–$1.2M in the first 12 months of implementation.