inventory-forecast
Installation
SKILL.md
Inventory Forecast
Running out of stock is one of the most costly mistakes in ecommerce — it tanks your search rankings, hands sales to competitors, and can take weeks to recover from. This skill projects inventory needs by analyzing historical order velocity, seasonal demand patterns, supplier lead times, and upcoming promotional events so you can place reorders with confidence before stockouts damage your momentum.
Quick Reference
| Decision | Strong | Acceptable | Weak |
|---|---|---|---|
| Velocity baseline | Trailing 90-day daily average, weighted toward recent 30 days, excluding promo spikes | Trailing 30-day simple average | Lifetime average or single best month |
| Safety stock | Statistical: z-score × demand std dev × √lead time | Fixed 20-30% buffer on lead-time demand | None, or "whatever is left over" |
| Lead time input | Full door-to-door: production + QC + freight + customs + receiving, with variance | Supplier's quoted production time + standard transit | Supplier's quoted production time alone |
| Promo demand lift | Modeled from your own past promo lift for the same SKU/category | Category benchmark (e.g., 2-4x for flash sales) | Guessing or ignoring the promo |
| Seasonality | Month-over-month index built from 12+ months of own data | Platform/category seasonal curve (e.g., Q4 lift) | Assuming flat demand year-round |
| Reorder trigger | Reorder point = (daily velocity × lead time) + safety stock, reviewed weekly | Fixed calendar reorders (e.g., monthly) sized by forecast | Reordering when stock "looks low" |
| Overstock handling | Flag SKUs with >90 days of cover and propose markdown/bundle plan | Note slow movers in the report | Treat all SKUs the same |