Total Invoiced Revenue
£18.36M
104 Calendar Weeks
Physical Volume Shipped
10.44M
Physical units
Active Forecast SKUs
1,760
Class A & B continuing
Full Catalog Split
1,044 A • 1,265 B
2,526 Class C long-tail
Total Warehouse Throughput Velocity (104 Weeks)
Weekly unit volume. Shaded red vertical zones indicate annual Christmas holiday shutdowns.
Pareto Revenue Split
Classic 80/20 commercial concentration
Class A: 80% Rev
Class B: 15% Rev
Class C: 5% Rev
SKU Trajectory & Non-Parametric Uncertainty Fan
Historical sales seamlessly blended into a 13-week forward trajectory bounded by empirical P10–P90 quantiles.
13-Week Future Replenishment Schedule
| Week Horizon | Calendar Start | Point Forecast (MA4) | P10 Lower Bound | P90 Upper Bound | 80% Uncertainty Band | Cumulative Trajectory |
|---|
Matched SKUs
1,760
Filtered item pool
Total Recommended ROP
498.4K
Reorder point trigger units
Quantile Safety Stock
208.6K
Dynamic empirical buffer
Classical 95% Safety Stock
244.3K
Gaussian formula buffer
Replenishment Parameter Catalog
Sizing comparison across all 1,760 SKUs. Click column headers to sort.
| StockCode | Description | ABC | Pattern | Lead Time | Mean Dem (μd) | Quantile SS | Classical 90% | Classical 95% | ROP (Rec) | Divergence |
|---|
Fold 3 Peak Holiday Replenishment Simulation
Simulated continuous replenishment on 1,636 active SKUs during Q4 autumn surge (1,311,074 actual units).
| Inventory Strategy | Fill Rate | Stockout Wks | Avg Overstock |
|---|
Corrected Strategic Policy Synthesis:
When evaluating the fair, matched-service-level comparison between Quantile ROP and Classical 90% CSL, both perform comparably in aggregate simulation (11.03% vs 13.89% stockouts; 79.20% vs 83.35% fill rate). Both decisively defeat the naive 4-week holding benchmark (19.75% stockout rate).
Quantile ROP is recommended structurally: it avoids forcing a symmetric Gaussian bell curve on retail demand where 36.3% of SKUs are intermittent or lumpy, eliminating cash lockup on slow movers while dynamically capturing seasonal tail acceleration.
SKU-Level Point Forecast Leaderboard
| Model Candidate | WAPE (%) | Bias (%) | MASE |
|---|
Macro Warehouse Aggregate Volume by Fold
| Backtest Horizon | Model | Macro WAPE | Macro Bias | Status |
|---|
Executive Discoveries & Strategic Roadmap
Synthesized takeaways from data cleaning, wholesale segregation, predictive backtesting, and inventory simulation.