inventory-optimization

name	inventory-optimization
description	Optimize inventory levels through safety stock calculation, reorder points, ABC/XYZ classification, carrying cost analysis, and stockout risk assessment. TRIGGER when: user says /inventory-optimization, "optimize inventory", "safety stock", "reorder point", "inventory analysis", "stock levels".
argument-hint	[product/SKU-group] [warehouse-or-location] [service-level-target]
user-invocable	true

Inventory Optimization

Right-size inventory across the network to maximize service levels while minimizing carrying costs, obsolescence risk, and working capital consumption.

---

Input Gathering

Input	Description	Required
SKU master data	Product list with unit cost, lead time, shelf life	Yes
Demand history	Minimum 12 months of consumption or sales data	Yes
Current stock positions	On-hand, in-transit, on-order by location	Yes
Target service level	Fill rate or cycle service level (e.g., 95%, 98%)	Yes
Lead time data	Supplier lead times with variability	Yes
Cost parameters	Holding cost %, ordering cost per PO, stockout penalty cost	No
Storage constraints	Warehouse capacity, shelf life limits, lot-size restrictions	No

---

Step-by-Step Process

Step 1 — ABC/XYZ Classification

Segment the portfolio on two dimensions:

ABC (Value) — based on annual revenue or cost of goods sold:

Class	Share of Revenue	Typical SKU %	Management Approach
A	~80%	~20%	Tight control, frequent review
B	~15%	~30%	Moderate control, periodic review
C	~5%	~50%	Simple rules, exception-based

XYZ (Variability) — based on coefficient of variation (CV) of demand:

Class	CV Range	Demand Pattern	Forecast Reliability
X	0 - 0.5	Stable	High
Y	0.5 - 1.0	Variable	Medium
Z	> 1.0	Erratic	Low

Create a 3x3 matrix (AX, AY, AZ, BX, BY, BZ, CX, CY, CZ) and assign inventory policies per cell.

Step 2 — Safety Stock Calculation

Calculate safety stock using the appropriate method for each segment:

Segment	Method	Formula
AX, BX	Statistical	SS = Z x sqrt(LT x sigma_d^2 + d_avg^2 x sigma_LT^2)
AY, BY	Statistical + buffer	SS = Z x sigma_demand_during_LT x (1 + buffer%)
AZ, BZ	Simulation / min-max	Run Monte Carlo on demand and lead time distributions
CX	Fixed weeks of supply	SS = avg_weekly_demand x weeks_target
CY, CZ	Min-max with high max	Set min = max observed demand during LT x 1.2

Where:

• Z = service level Z-score (e.g., 1.65 for 95%, 2.05 for 98%)
• LT = average lead time
• sigma_d = standard deviation of demand
• sigma_LT = standard deviation of lead time

Step 3 — Reorder Point and Order Quantity

• Reorder Point (ROP) = Average demand during lead time + Safety stock.
• Economic Order Quantity (EOQ) = sqrt(2 x D x S / H), where D = annual demand, S = ordering cost, H = holding cost per unit.
• Apply practical constraints: minimum order quantity (MOQ), pack sizes, container utilization, shelf-life limits.
• For expensive A-items, consider lot-for-lot ordering to minimize holding cost.

Step 4 — Carrying Cost Analysis

Build a full cost-to-carry model:

Cost Component	Typical Range	Calculation Basis
Capital / opportunity	8-15%	WACC x average inventory value
Warehousing	2-5%	Space, labor, utilities
Insurance	0.5-1%	Insured value
Obsolescence / shrinkage	2-10%	Historical write-off rate
Total carrying cost	15-30%	Sum of above

Compare current vs. proposed inventory investment and quantify savings.

Step 5 — Stockout Risk Assessment

For each SKU-location, calculate:

• Probability of stockout per cycle = 1 - Service Level achieved.
• Expected stockout units = integral of demand distribution beyond ROP during lead time.
• Revenue at risk = Expected stockout units x unit price x (1 - substitution rate).
• Customer impact score = Revenue at risk x customer criticality weight.

Rank SKUs by customer impact score and flag the top decile for immediate action.

Step 6 — Policy Recommendations and Monitoring

• Set inventory policies (ROP, safety stock, review period) per ABC/XYZ segment.
• Define exception triggers: days of supply < threshold, excess > X weeks, slow-moving flag.
• Create a dashboard cadence: A-items reviewed weekly, B-items biweekly, C-items monthly.
• Establish an obsolescence review quarterly for items with < 60 days shelf life remaining or > 12 months of supply.

---

Output Format

## Inventory Optimization Report — [Scope] — [Date]

### Executive Summary
- Total SKUs analyzed: [count]
- Current inventory value: [$X]
- Recommended inventory value: [$Y]
- Projected savings: [$Z] ([%] reduction)
- Service level impact: [current] -> [projected]

### ABC/XYZ Classification Summary
| Segment | SKU Count | Inventory Value | Revenue Share | Recommended Policy |
|---------|-----------|-----------------|---------------|--------------------|
| AX      | ...       | ...             | ...           | ...                |
| AY      | ...       | ...             | ...           | ...                |
| ...     | ...       | ...             | ...           | ...                |

### Safety Stock and Reorder Points (Top SKUs)
| SKU | Current SS | Proposed SS | ROP | EOQ | Days of Supply |
|-----|-----------|-------------|-----|-----|----------------|
| ... | ...       | ...         | ... | ... | ...            |

### Stockout Risk Register
| SKU | Stockout Probability | Revenue at Risk | Impact Score | Action |
|-----|----------------------|-----------------|--------------|--------|
| ... | ...                  | ...             | ...          | ...    |

### Recommendations
1. [Specific, actionable recommendations]
2. ...

### Implementation Roadmap
- [Week 1-2]: ...
- [Week 3-4]: ...

---

Quality Checklist

• Demand data covers at least 12 months and is cleansed of stockout periods
• ABC/XYZ classification completed and validated with stakeholders
• Safety stock method appropriate for each segment's demand pattern
• Lead time variability captured (not just averages)
• Carrying cost rate reflects actual company cost structure
• Stockout cost includes both direct and downstream impact
• MOQ, lot size, and shelf-life constraints applied to order quantities
• Exception triggers defined for proactive monitoring
• Recommendations quantified with expected financial impact
• Review cadence established per segment

---

Edge Cases

• New SKU with no demand history: Use analogous product demand profile; set conservative safety stock at 1.5x expected lead time demand until 3 months of actuals accumulate.
• Highly seasonal products: Calculate safety stock using within-season variability only; pre-build inventory using seasonal index.
• Perishable goods with short shelf life: Add shelf-life constraint to EOQ; reduce max stock to shelf-life minus lead time; implement FEFO (First Expired, First Out).
• Long and variable lead times (e.g., overseas sourcing): Use lead-time demand distribution from simulation; consider dual sourcing to reduce effective variability.
• Vendor-managed inventory (VMI): Shift analysis to consumption-based replenishment; share demand signals directly with supplier.
• Promotional spikes: Separate base and promotional demand; plan promotional inventory as a one-time forward buy outside the standard safety stock framework.