safety-stock-reviewlisted
Install: claude install-skill gulmezeren2-byte/industrial-engineering-ai-skills
# Safety Stock Review
The textbook formula assumes roughly normal demand. Real portfolios contain SKUs where that assumption fails badly - the skill's job is to compute the number AND say where it can be trusted.
## Required data
Per-SKU demand history (`sku`, `period`, `qty`, 12+ periods), lead time (with variability if available), and the service target. Clarify early **which service the target means**: cycle service (probability of no stockout per cycle) or fill rate (share of units served) - contracts usually mean fill rate, formulas usually compute cycle service.
## Workflow
1. **Classify first.** Compute CV and zero-period share per SKU. For CV >= 1.0 or intermittent demand, state up front that the normal-formula result will be optimistic.
2. **Compute the formula result:** SS = z * sigma_d * sqrt(LT), ROP = mu_d * LT + SS (demand-period units consistent with LT). If lead time varies, use the extended form with the sigma_LT term - ignoring lead-time variance is the most common silent understatement.
3. **Stress-test empirically.** Set stock at mu + SS and replay the actual history: report both achieved cycle service (share of periods fully covered) and achieved fill rate (units served / units demanded). Zero-demand periods pass cycle service for free - fill rate is the honest one on intermittent items.
4. **Show the cost of nines.** SS at 90/95/98/99% targets for the SKUs in question - service targets are pricing decisions, and the curve makes that visible.
5. **Re