The Challenge
Volatile demand patterns and poor supplier visibility led to chronic stockouts on key SKUs and costly overstocking of slow-moving products, tying up capital and warehouse space.
Our Solution
We implemented ensemble ML models trained on 5 years of sales data, integrating external signals like seasonality, economic indicators, and supplier lead times. An automated reorder engine acts on model outputs in real time.
Results
- 62% reduction in stockout events
- 28% decrease in excess inventory carrying costs
- Supplier lead time visibility improved across 200 vendors
- Demand forecast accuracy improved from 71% to 94%
- Full ROI achieved in under 4 months