Week 2 | Session 4: Analytical Product Segmentation & Kraljic Matrix

Course: Supply Chain Digitization

---

Session Context

Note

Previous sessions covered: SC challenges → segmentation strategies and how to match them to SC type.

This session: A deep dive into Product Segmentation using two analytical frameworks:

1. Demand Variability vs. Weekly Sales Quadrant - a data-driven scatter-and-classify method
2. Kraljic Matrix - a sourcing and procurement decision framework

Both are practical tools. Expect case-based exam questions where you must classify SKUs or components and recommend SC or procurement strategies.

---

Part 1 - Analytical Product Segmentation: Demand Variability & Weekly Sales

Data-Driven Segmentation

Case Setup

Company: Electronic gadgets manufacturer with a large variety of SKUs. Products: Smartphones, tablets, smart watches, and other gadgets. Operations: 2 cities - City A and City B; 2 channels - Online and Retail. Data available: Weekly sales for 20 SKUs over 8 weeks for both cities. Annual revenue: USD 1 billion.

Segmentation variables used:

Variable	What It Measures
Average Weekly Sales	Volume indicator - how much of a SKU is sold per week on average
Demand Variability / Order Variability	Stability indicator - how much demand fluctuates from week to week

---

Measuring Demand Variability - Coefficient of Variation (CoV)

Note

Demand variability = fluctuations in orders, including changes in the number and size of orders over time.

Tool used: Coefficient of Variation (CoV)

Why CoV and not just Standard Deviation?

Measure	What It Captures	Limitation
Standard Deviation (σ)	Absolute variability around the mean - the raw fluctuation level	Cannot be compared across SKUs with different demand levels
CoV (σ / μ)	Variability relative to average demand - normalised across SKUs	None - this is precisely why it is preferred

CoV allows fair comparison across SKUs with very different demand levels - a major analytical advantage over using σ alone.

Formulas:

Note

CoV = σ (Standard Deviation) / μ (Mean)

Mean (μ) = Σxᵢ / n → Sum of all order quantities divided by the number of observations

Standard Deviation (σ) = √[ Σ(xᵢ − x̄)² / n ] → where xᵢ = individual order quantity, x̄ = average demand, n = number of observations

Tip

Interpreting CoV: CoV ≈ 0.1 → very stable demand. CoV ≥ 0.5 → high variability. Always use CoV to compare across SKUs, not raw σ.

---

Steps to Calculate & Segment (Excel-Based Workflow)

1. Collect data - gather 8-week weekly demand data for all 20 SKUs for City A and City B separately
2. Calculate SKU-wise average demand - use the AVERAGEIFS conditional formula for each city
3. Calculate SKU-wise standard deviation - use the STDEVIFS conditional formula for each city
4. Compute CoV - divide σ by μ for each SKU per city: CoV = σ / μ
5. Plot a scatter - X-axis: CoV (demand variability); Y-axis: Average weekly sales - one scatter per city
6. Combine City A + City B demand - compute the joint average demand, joint σ, and joint CoV for each SKU
7. Plot the combined (joint) scatter - this represents the Distribution Centre view serving both cities
8. Build the 4-Quadrant Matrix - assign an SC strategy to each product group based on quadrant position

Note

★ Key point: The combined (joint) CoV represents the total demand variability faced by the distribution centre serving both cities simultaneously - this is the operationally relevant view for SC design.

---

Raw Weekly Sales Data Table

Structure of the raw dataset (sample rows shown):

Week	SKU	City	Units Sold
1	SKU13	City A	105
2	SKU7	City A	144
3	SKU15	City A	93
4	SKU12	City B	136
5	SKU1	City B	133
6	SKU15	City A	65
7	SKU4	City A	198
8	SKU13	City B	194
…	…	…	…

Note

Full dataset = 20 SKUs × 8 weeks × 2 cities. The SKU-level averages, standard deviations, and CoV computed from this raw table are summarised below.

---

CoV Calculations - City A

Product	Average Demand	Std. Deviation	CoV
SKU1	117.2	38.5	0.3
SKU2	117.7	40.2	0.3
SKU3	129.7	36.3	0.3
SKU4	129.6	41.8	0.3
SKU5	85.3	43.9	0.5
SKU6	166.8	24.7	0.1
SKU7	139.6	39.8	0.3
SKU8	117.4	47.9	0.4
SKU9	108.0	40.6	0.4
SKU10	134.6	44.4	0.3
SKU11	142.8	50.7	0.4
SKU12	130.2	56.8	0.4
SKU13	99.0	25.6	0.3
SKU14	139.3	41.4	0.3
SKU15	92.8	47.0	0.5
SKU16	129.1	47.9	0.4
SKU17	100.3	46.4	0.5
SKU18	118.6	45.6	0.4
SKU19	127.7	29.3	0.2
SKU20	117.9	49.8	0.4

---

CoV Calculations - City B

Product	Average Demand	Std. Deviation	CoV
SKU1	130.9	49.9	0.4
SKU2	148.0	36.4	0.2
SKU3	127.2	37.4	0.3
SKU4	128.4	40.7	0.3
SKU5	122.3	62.1	0.5
SKU6	126.8	48.9	0.4
SKU7	133.5	45.3	0.3
SKU8	104.9	38.1	0.4
SKU9	102.0	33.6	0.3
SKU10	127.6	52.0	0.4
SKU11	104.9	43.2	0.4
SKU12	117.4	31.2	0.3
SKU13	124.6	60.2	0.5
SKU14	154.7	25.6	0.2
SKU15	106.5	49.9	0.5
SKU16	110.7	47.0	0.4
SKU17	113.1	38.2	0.3
SKU18	142.7	31.0	0.2
SKU19	126.7	42.7	0.3
SKU20	113.5	31.2	0.3

---

Joint (Combined) CoV - Distribution Centre View

Product	Average Demand	Std. Deviation	CoV
SKU1	125.8	45.1	0.4
SKU2	136.8	39.7	0.3
SKU3	128.4	35.2	0.3
SKU4	129.1	39.8	0.3
SKU5	107.5	56.1	0.5
SKU6	140.1	45.7	0.3
SKU7	137.5	40.5	0.3
SKU8	110.7	41.9	0.4
SKU9	105.0	35.9	0.3
SKU10	131.9	46.5	0.4
SKU11	115.7	46.9	0.4
SKU12	120.9	38.7	0.3
SKU13	109.7	43.0	0.4
SKU14	147.8	33.6	0.2
SKU15	99.7	46.7	0.5
SKU16	120.4	47.1	0.4
SKU17	105.4	42.7	0.4
SKU18	129.1	40.6	0.3
SKU19	127.2	34.9	0.3
SKU20	115.8	40.9	0.4

---

Scatter Plot - City A & City B Combined

Note

Plotting each SKU as CoV (x-axis) against average weekly sales (y-axis) - separately for City A and City B - shows the same broad pattern in both cities: higher demand variability tends to accompany lower average sales. The two cities are then combined into the joint distribution-centre view below, which is the operationally relevant scatter for SC design.

---

Scatter Plot - Distribution Centre (Joint) View

Distribution-centre view - each point is one of the 20 SKUs, positioned by its joint demand variability (CoV) and average weekly sales. The downward drift confirms that the most variable SKUs are generally the lower-selling ones.

---

4-Quadrant Strategy Matrix

The scatter plots are divided into four quadrants based on threshold values for CoV and average weekly sales. Each quadrant maps to a distinct SC strategy.

↑ High average weekly sales

Q1 · Push (low CoV, high sales)

• Essential / high-volume products
• Stable, predictable → Efficient SC

Q3 · Hybrid push-pull (high CoV, high sales)

• Consumer electronics / high-tech
• High volume and high variability - most complex

Q4 · Hybrid push-pull (low CoV, low sales)

• Basic / slow-moving products
• Combination set by product requirements

Q2 · Pull (high CoV, low sales)

• Customised / niche products
• Low volume + unpredictable → Responsive SC

Low demand variability (CoV) ← → High demand variability (CoV)  ·  ↓ Low average weekly sales

The scatter is split into four quadrants by thresholds on CoV and average weekly sales; each quadrant maps to a distinct supply-chain strategy.

Quadrant 1 - Low CoV + High Sales (Top-Left)

Push Strategy

Attribute	Detail
Product type	Essential / High-volume products
Characteristics	Always in demand, stable, predictable
Examples	Standard commodity goods, staple consumer products
SC type	Efficient Supply Chain
Strategy	PUSH - forecast-driven, mass production

Quadrant 2 - High CoV + Low Sales (Bottom-Right)

Pull Strategy

Attribute	Detail
Product type	Customised / Niche products
Characteristics	Low demand volume + high unpredictability
Examples	Special-edition collectibles, highly customised products
SC type	Responsive Supply Chain
Strategy	PULL - demand-driven, high customisation

Quadrant 3 - High CoV + High Sales (Top-Right)

Hybrid Push-Pull

Attribute	Detail
Product type	Consumer electronics / High-tech products
Characteristics	High volume AND high variability - the most complex quadrant to manage
Examples	Consumer electronics, tech gadgets with rapid model changes
Strategy	Hybrid Push-Pull - apply push-pull boundary framework from Session 3; boundary position depends on level of customisation required

Quadrant 4 - Low CoV + Low Sales (Bottom-Left)

Hybrid Push-Pull

Attribute	Detail
Product type	Basic / Slow-moving products
Characteristics	Low sales + stable demand - neither high priority nor high risk
Examples	Non-seasonal apparel basics, basic electronic components
Strategy	Hybrid Push-Pull - specific combination determined by individual product requirements

Note

★ Quick Recall:

• Q1 (Low CoV, High Sales) = Push
• Q2 (High CoV, Low Sales) = Pull
• Q3 (High CoV, High Sales) = Hybrid Push-Pull
• Q4 (Low CoV, Low Sales) = Hybrid Push-Pull

---

Part 2 - Kraljic Matrix: Procurement & Sourcing Segmentation

Sourcing Strategy Framework

Note

Framework by: Peter Kraljic

Purpose: Classify the components and inputs required for manufacturing based on two strategic factors, and use that classification to determine the right supplier relationship strategy and procurement approach for each component type.

Two Axes of the Kraljic Matrix

Axis	What It Measures	Key Factors
X-axis - Supply Risk	How difficult or risky it is to procure the component	Availability of the component; number of suppliers; competitive demand; make-or-buy feasibility; other supply-side risk factors
Y-axis - Profit Impact	How significantly the component affects business profitability	Volume required; percentage of total purchase cost; direct or indirect impact on business growth

---

4 Quadrants of the Kraljic Matrix

↑ High profit impact

Leverage (low risk, high impact)

• Abundant supply, strong buyer power
• Exploit purchasing power: competitive bidding, hard negotiation

Strategic (high risk, high impact)

• Critical, hard to switch suppliers
• Long-term partnerships, collaboration, secure continuity

Non-critical (low risk, low impact)

• Standardised, many suppliers
• Streamline & automate procurement

Bottleneck (high risk, low impact)

• Low control, but can disrupt operations
• Secure supply: safety stock, alternate sourcing, monitor

Low supply risk ← → High supply risk  ·  ↓ Low profit impact

The Kraljic matrix classifies purchased components by supply risk and profit impact, assigning each of the four quadrants a distinct procurement strategy.

1. Non-Critical Items - Low Supply Risk + Low Profit Impact

Attribute	Detail
Characteristics	Highly standardised; abundant supply - many suppliers available; easy to manage
Strategy	Streamline procurement - reduce administrative effort, automate ordering where possible

2. Leverage Items - Low Supply Risk + High Profit Impact

Attribute	Detail
Characteristics	Abundant supply available but high impact on profit; company holds strong purchasing power
Strategy	Exploit full purchasing power - competitive bidding, aggressive price negotiation, good pricing strategy

3. Strategic Items - High Supply Risk + High Profit Impact

Attribute	Detail
Characteristics	Most critical category - high risk AND high profit impact; affects the business on a long-term basis; cannot switch suppliers easily
Strategy	Long-term supplier relationships + deep collaboration - partnership approach, joint development, strategic alliances; focus on supply continuity

4. Bottleneck Items - High Supply Risk + Low Profit Impact

Attribute	Detail
Characteristics	High supply risk but minimal direct profit impact; very low control over suppliers; unavailability can still disrupt operations even when profit impact is low
Strategy	Ensure availability through safety stock and alternate sourcing; monitor closely

---

Kraljic Matrix - Quick Reference

Quadrant	Supply Risk	Profit Impact	Procurement Strategy
Non-Critical	Low	Low	Streamline and automate procurement
Leverage	Low	High	Negotiate hard; competitive bidding
Strategic	High	High	Long-term partnerships; deep collaboration
Bottleneck	High	Low	Secure supply; hold safety stock; monitor closely

Tip

★ Exam Tip - Kraljic Matrix: Classify by Supply Risk (X-axis) and Profit Impact (Y-axis). Four quadrants = Non-Critical, Leverage, Strategic, Bottleneck. Each requires a fundamentally different supplier strategy.

Watch out for Bottleneck items: Low profit impact does not mean low priority - the disruption risk to operations makes them critical to manage actively despite their limited financial significance.

---

Session Summary - Two Analytical Tools

Tool	Axes	Output	Strategy Assigned
Demand Variability vs. Weekly Sales Quadrant	X: CoV (variability), Y: Avg weekly sales	4-quadrant product classification	Q1 = Push | Q2 = Pull | Q3 & Q4 = Hybrid Push-Pull
Kraljic Matrix	X: Supply risk, Y: Profit impact	4-quadrant component classification	Non-Critical = Streamline | Leverage = Negotiate | Strategic = Partner | Bottleneck = Secure

Note

Both frameworks are data-driven and directly actionable - they take measurable inputs (sales data, supply risk, profit contribution) and produce clear strategic recommendations. In exams, you will likely be given a dataset or scenario and asked to classify items and justify the recommended strategy using one or both frameworks.