Finding the inventory hiding in plain sight
A decision-oriented segmentation model for finding excess inventory without putting customer service at risk.
Problem
Aggregate inventory was rising, but the existing dashboard could not distinguish strategic buffers from avoidable excess.
Background
Different demand patterns require different policies. A single days-of-cover threshold created noise and eroded trust.
Methodology
I combined ABC value bands, XYZ variability, demand recency, lead time, open supply, and target service level into explainable exception rules.
Analysis
The model prioritised value at risk while preserving the operational context needed by planners.
Results
The outcome was not another passive dashboard. It was a ranked work queue with reasons, suggested actions, owners, and review dates.
Lessons learned
Explainability drives adoption. A slightly simpler model that a planner can challenge often creates more value than an opaque optimisation score.