Segmentation-Driven Inventory Policy

Definition

Segmentation-driven inventory policy assigns differentiated replenishment rules, service level targets, safety stock methods, and review cadences to groups of SKUs based on their economic value (ABC) and demand variability (XYZ). The goal is to concentrate working capital and planner attention where stockouts are most costly, while reducing overhead on low-value or unpredictable items.

Managing every SKU with a single policy wastes capital on C-items and under-serves A-items. Segmentation creates policy tiers that match inventory investment to business risk.

---

ABC Classification (Value Tier)

Rank SKUs by annual consumption value (unit cost × annual volume) or annual margin contribution:

Tier	Typical share of SKUs	Share of total value	Policy intent
A	10–20%	70–80%	Maximum service; tight control; frequent review
B	30–40%	15–25%	Balanced service; standard policy
C	40–50%	5–10%	Lean inventory; automated replenishment; consider MTO

Use margin contribution (not just revenue) for A/B/C when SKU profitability varies significantly.

---

XYZ Classification (Demand Variability Tier)

Measure coefficient of variation (CoV = standard deviation / mean) of demand over a rolling 12-month window:

Tier	CoV threshold	Demand pattern	Forecast reliability
X	CoV < 0.5	Stable, predictable	High — statistical safety stock works well
Y	0.5 ≤ CoV < 1.0	Seasonal or trend-driven fluctuation	Moderate — buffer stocks required
Z	CoV ≥ 1.0	Erratic, lumpy, or intermittent	Low — statistical models unreliable

Recalculate CoV quarterly. SKUs migrate between tiers as demand patterns shift.

---

9-Box Policy Matrix

	X (stable)	Y (seasonal)	Z (erratic)
A (high value)	AX: 98% fill rate; CSL safety stock; continuous review; MTS	AY: 96% fill rate; CSL + lead-time buffer; monthly review; MTS	AZ: 94% fill rate; Croston method or days-of-cover buffer; quarterly review; MTS with elevated SS
B (mid value)	BX: 95% fill rate; statistical SS; periodic review (2-week); MTS	BY: 92% fill rate; buffer stock; periodic review (2-week); MTS	BZ: 90% fill rate; elevated buffer; monthly review; consider MTO for low-volume variants
C (low value)	CX: 90% fill rate; lean SS; automated replenishment; MTS	CY: 88% fill rate; moderate buffer; periodic review (monthly); assess MTO	CZ: 85% or MTO; minimal safety stock; consider consignment, delisting, or make-to-order

[!note] These targets are starting-point benchmarks, not universal rules. Calibrate to your stockout cost, carrying cost rate, and customer contract requirements.

---

MTO vs MTS Decision Rules

Make-to-Stock (MTS): Hold finished goods inventory; replenish to a target level. Appropriate when:

• Demand is frequent (>4 orders/year per SKU)
• Lead time exceeds customer-acceptable wait time
• SKU is in AX, AY, AZ, BX, BY tiers
• Stockout cost (lost sale, penalty, substitution cost) > carrying cost

Make-to-Order (MTO): Produce or procure only against confirmed orders. Appropriate when:

• Demand is infrequent or lumpy (Z-items with <4 orders/year)
• SKU is configurable or highly perishable
• Customer lead time tolerance exceeds production/procurement lead time
• Carrying cost > expected stockout cost (typical for CZ items)

Hybrid / Postponement: Stock a common semi-finished platform (MTS); finish to order. Used for BY/BZ items with high SKU count but shared upstream components.

---

Service Level Differentiation by Segment

Segment	Cycle service level	Safety stock method
AX	98–99%	z × σ_demand × √LT (standard normal)
AY	96–97%	z × σ_LTD (lead-time demand std dev) + seasonal buffer
AZ	93–95%	Croston’s method for intermittent demand; or days-of-cover (4–8 weeks)
BX	94–96%	z × σ_demand × √LT
BY	91–93%	σ_LTD buffer
BZ	88–91%	Days-of-cover (2–4 weeks); evaluate MTO
CX	88–92%	Lean SS; automated min/max
CY	85–90%	Minimal buffer; periodic review
CZ	≤85% or MTO	Near-zero SS; order-driven only

Z-scores for common service levels: 90% → z=1.28; 95% → z=1.65; 97% → z=1.88; 98% → z=2.05; 99% → z=2.33

---

Review Policy by Segment

Segment	Review type	Review frequency	Reorder trigger
AX, AY	Continuous (s,S)	Real-time	Drop below reorder point s
AZ	Periodic + exception	Weekly	Drop below days-of-cover threshold
BX, BY	Periodic (R,s,S)	Bi-weekly	Scheduled review + reorder point check
BZ	Periodic	Monthly	Scheduled review
CX	Automated min/max	Monthly or event-driven	System-generated PO when below min
CY, CZ	Periodic or MTO	Monthly or quarterly	Demand signal or manual review

Continuous review (triggered by every transaction) is cost-justified for A-items; periodic review (scheduled intervals) is sufficient for B and C items and reduces planner workload.

---

Safety Stock by Segment

Safety stock absorbs two sources of uncertainty: demand variability and supply lead time variability.

Standard formula: SS = z × σ_LTD where σ_LTD = √(LT × σ_d² + d̄² × σ_LT²)

Segment	Primary driver	Practical approach
AX	Lead time variability > demand variability (stable demand)	Calculate σ_LTD rigorously; review quarterly
AY	Both demand and lead time variability	Full σ_LTD formula; add seasonal factor
AZ	Demand is intermittent — normal distribution doesn’t apply	Croston’s method or empirical days-of-cover
BX–BY	Simplified σ_LTD acceptable	Use standard formula; relax to bi-weekly review
CX–CZ	Working capital cost often > safety benefit	Min-max policy or MTO; avoid over-investing

---

Implementation Steps

1. Data preparation: Pull 12–24 months of demand history by SKU. Cleanse promotions, stockout-driven zero-demand periods, and one-time spikes.
2. ABC classification: Rank by annual consumption value. Set A/B/C boundaries at 80/95/100% of cumulative value (adjust if top decile is too concentrated).
3. XYZ classification: Calculate CoV per SKU. Assign X/Y/Z thresholds. Validate Z-items are genuinely erratic, not data quality issues.
4. Policy design: Assign service level targets and SS methods per the 9-box matrix. Model working capital impact before finalizing.
5. Planner workload test: Verify A-item count is manageable (typically <200 SKUs per planner for continuous review).
6. System configuration: Enter reorder points, safety stock quantities, and review cycles into ERP/WMS or planning system.
7. Governance: Review SKU segment assignments quarterly. Flag items that migrate from AX → AZ (demand pattern shift) for root cause investigation.