$390M in plan overshoot across 16 quarters. Forecasting at item-color-size granularity revealed $75M in conservative annual opportunity.
The planning team's weighted MAPE was 28% at the item-color level, with a systematic bias toward over-forecasting: $390M in plan overshoot versus $110M in undershoot across 16 quarters. True stockout rate reached about 5%, representing $6M in annual lost demand. MOQ constraints were forcing excess purchasing on approximately 900 products. Size allocation errors averaged 3 percentage points per size across 150,000 records, a seemingly small number that compounds significantly at assortment scale. Annual plan overshoot totaled $99M.
Profitmind conducted a comprehensive proof of concept analyzing over 5million demand records across all divisions from FY2021 through FY2025. The AI forecasted at item-color-size level — 14,000 combinations versus the client's existing 3,000 — and at weekly frequency versus quarterly. Results were validated at three tiers: Top 10% of the assortment, Top 50%, and a full assortment extrapolation. All figures were hard-dollar only; no soft benefits were included in the analysis.
Conservative annual impact across the full assortment reached $75M — equivalent to 12% of annual revenue. Size allocation analysis alone drove $13M in revenue opportunity and freed $2M in working capital, value that is only visible when forecasting at item-color-size granularity. For the top 50% of the assortment, annual revenue upside reached $28M and working capital savings reached $27M. The forecast accuracy improvement of 21% relative to the existing plan translates directly to reduced overbuy, fewer markdowns, and better in-stock rates on productive inventory.