🚚 Supply Chain & Logistics · Planning & Inventory

Build a demand baseline you can beat

Cleanse history, segment by variability, and benchmark naive forecasts before fancy models.

intermediate~35 minSupply PlannersLogistics ManagersProcurement

Steps

1Cleanse history: mark stockouts, promotions, one-time events explicitly
2Segment SKUs by ABC volume × XYZ variability — different methods per class
3Benchmark naive (last period / same-period-last-year) first, always
4Only adopt complex methods that beat naive on YOUR data out-of-sample
5Forecast at the grain decisions are made; aggregate for reporting only
6Track bias separately from accuracy; systematic over/under is fixable

Common Pitfalls

▲Promotional spikes trained into base demand forever
▲MAPE worship while every forecast runs systematically high

Commands

Install with skills CLI

$ npx skills add aniruddhaadak80/skills --skill planning-inventory-demand-baseline-forecast

Install globally

$ npx skills add aniruddhaadak80/skills --skill planning-inventory-demand-baseline-forecast -g

Tags

#forecasting#demand-planning#supply-chain-logistics#planning-inventory

Install this skill

$ npx skills add aniruddhaadak80/skills --skill planning-inventory-demand-baseline-forecast -a opencode

SKILL.md

# Build a demand baseline you can beat

> Cleanse history, segment by variability, and benchmark naive forecasts before fancy models.

## When to Use This Skill

Cleanse history, segment by variability, and benchmark naive forecasts before fancy models.

## Steps

1. Cleanse history: mark stockouts, promotions, one-time events explicitly
2. Segment SKUs by ABC volume × XYZ variability — different methods per class
3. Benchmark naive (last period / same-period-last-year) first, always
4. Only adopt complex methods that beat naive on YOUR data out-of-sample
5. Forecast at the grain decisions are made; aggregate for reporting only
6. Track bias separately from accuracy; systematic over/under is fixable

## Common Pitfalls

- Promotional spikes trained into base demand forever
- MAPE worship while every forecast runs systematically high

View source on GitHub ↗

Statistical buffers per SKU-class matching stated service levels and lead-time truth.

🚚 Supply Chain & Logistics·~30m

Open →

Write a disruption playbook before you need itadvanced

Pre-ranked responses to supplier/port/demand shocks with decision authority mapped.

🚚 Supply Chain & Logistics·~40m