Chapter 2 of 4

Network Design and Sourcing Strategy

3 sections · ~5 min read · 3 practice questions · part of Supply Chain Strategy

Learning objectives

  • Distinguish structural supply chain decisions (plant count, location, capacity) from operating decisions (production schedule, transportation mode choice).
  • Compare sourcing models—in-house, single-source, dual-source, near-source—on cost, risk, and lead time.
  • Compute a simple total-cost comparison between two network options and identify the dominant cost driver.

Structural versus Operating Decisions

Supply chain decisions live on a spectrum of reversibility. Structural decisions—how many plants, where, of what capacity, what sourcing model—are expensive and slow to reverse; they lock in fixed costs and lead-time geometry for years. Operating decisions—production schedules within a plant, mode choice per shipment, safety-stock levels—are reversible within weeks or months. Many practitioners spend most of their time on operating decisions because they are visible and frequent, but the leverage is in structural decisions. A wrong network choice can absorb millions in avoidable cost and create lead-time exposure that no operating improvement can fix. A useful heuristic: if the decision will still be in force five years from now, treat it as structural and apply a heavier analytical lens—including sensitivity analysis on demand, FX, and labor rates—even when the time pressure is high. Reversibility also determines the appropriate modeling depth. Structural decisions justify network-optimization models with full cost breakdowns; operating decisions are often best left to local optimization within the structure.

Sourcing Models and Their Tradeoffs

Four sourcing models dominate: in-house, single-source, dual-source, and near-source. In-house maximizes control and captures margin but ties capital to operations and limits geographic reach. Single-source captures scale and deepens collaboration with the supplier but creates concentration risk and reduces negotiation leverage. Dual-source reduces single-point-of-failure exposure and supports supplier benchmarking, but raises complexity, splits volumes, and may reduce per-supplier scale economies. Near-source reduces lead time and certain geopolitical risks, at the cost of higher unit cost relative to low-cost regions. The right model depends on the part's strategic importance: commodity items tolerate single-source; engineered or sole-designed items benefit from dual-source qualification; mission-critical or rapidly evolving items may justify near-shoring or in-house. A common failure is applying the same sourcing model across the entire spend portfolio. The result is either excessive complexity (dual-source on commodity fasteners) or excessive risk (single-source on a sole-designed subassembly). Sourcing strategy should be tiered by part criticality, not by the purchasing department's preferred posture.

Total Cost of Network Choices

Comparing two network options requires more than unit price. Total cost includes fixed operating cost of facilities, variable production cost, transportation cost (including mode-dependent transit time and reliability), inventory carrying cost on the pipeline, customs and duty where applicable, and partner oversight or coordination cost. A common mistake is to compare only variable cost and ignore the inventory carrying cost created by longer lead times. A 28-day ocean pipeline vs a 5-day truck pipeline looks cheap on freight rate but carries a much larger in-transit and buffer inventory that may cost more than the freight savings. Working-capital cost is real cost and should be modeled with an explicit cost-of-capital rate and average pipeline inventory. Sensitivity analysis matters: model the decision under low, base, and high demand; under currency moves; under partner-performance scenarios. The structural decision that wins on every scenario is preferred; the one that wins only on base case is fragile. If no option wins across scenarios, prefer the one whose downside is smaller.

Worked example

Problem

Cascadia Tools is deciding how to serve its European market with annual demand of 220,000 units. Option A: ship from a single plant in the central US, ocean 28 days, $4.20/unit landed cost, plant fixed operating cost $5,000,000/year. Option B: contract manufacture in Eastern Europe, 5 days lead time, $6.80/unit landed cost, partner fixed cost $800,000/year plus oversight $400,000/year. Compare annual total cost and lead-time exposure.

Step by step

  1. Option A variable cost = 220,000 × $4.20 = $924,000. Plus fixed $5,000,000. Total = $5,924,000/year. Lead time 28 days.
  2. Option B variable cost = 220,000 × $6.80 = $1,496,000. Fixed = $800,000 + $400,000 = $1,200,000. Total = $2,696,000/year. Lead time 5 days.
  3. Variable difference: $1,496,000 − $924,000 = $572,000 more for B.
  4. Fixed difference: $5,000,000 − $1,200,000 = $3,800,000 less for B.
  5. Net annual saving of B over A = $3,800,000 − $572,000 = $3,228,000.
  6. Pipeline effect, computed separately so it is not double-counted: Option A holds 220,000 × (28/365) = 16,877 units in transit at $4.20 = $70,883 of working capital. Option B holds 220,000 × (5/365) = 3,014 units at $6.80 = $20,493. Average pipeline inventory falls by 13,863 units and working capital by about $50,390.
  7. At a 10% cost of capital that pipeline release is worth roughly $5,040 per year, which is real but trivial next to the $3.2M fixed-cost difference. Worth noting explicitly: this is a case where the working-capital argument, usually the one everybody forgets, is not the argument that decides it.

Answer. Option B costs $2,696,000 against $5,924,000, saving $3,228,000 per year and cutting lead time by 23 days, with a further $50,390 of working capital released from the pipeline. The dominant driver is fixed cost, not unit cost and not inventory. One caveat determines whether the comparison holds at all: the $5,000,000 plant overhead only counts if it is genuinely avoidable. If the US plant also serves other markets and would keep running without the European volume, the relevant figure is the incremental fixed cost of that volume, which could be near zero and would reverse the answer entirely. Establish which of those two situations you are in before presenting this analysis. Option B also introduces supplier-performance and oversight risk that the cost comparison does not price.

Practice

Work each question before opening the solution.

  1. What is the defining characteristic of a structural supply chain decision?

    Show solution for question 1

    It is expensive and slow to reverse, locking in cost and capability for years. Examples include plant location, number of facilities, capacity additions, and the fundamental sourcing model for a part.

  2. Why might a company choose dual-source even when single-source offers a lower unit price?

    Show solution for question 2

    Dual-source reduces single-point-of-failure risk, preserves negotiation leverage, supports supplier benchmarking, and protects against capacity or quality shocks at the primary supplier.

  3. What hidden cost most often distorts a network comparison?

    Show solution for question 3

    Working capital tied up in the longer pipeline. A cheaper but slower network option may carry so much in-transit and buffer inventory that the implied cost-of-capital charge exceeds the unit-price savings.