Facility Location and Network Design

Introduction

Facility location is the strategic decision concerning where an organisation should place plants, warehouses, distribution centres, service units, data centres, or other operating assets. Network design extends this decision by determining how facilities, suppliers, markets, transport links, inventories, and information flows should be configured as an integrated system. These decisions are among the most capital-intensive and difficult to reverse in operations management. A location may remain in use for decades, while its consequences influence cost, responsiveness, resilience, employment, environmental performance, and competitive advantage.

Location decisions are not simply a search for the cheapest land. They require evaluation of demand, transportation, labour, taxation, utilities, regulation, risk, sustainability, technology, and strategic fit. A low-cost location may be unattractive if it has unreliable electricity, weak supplier ecosystems, exposure to floods, or poor access to skilled workers. Conversely, a higher-cost location may be justified by superior service, innovation capability, market access, or resilience.

At postgraduate level, location analysis should be viewed as a multi-period and multi-objective decision under uncertainty. Quantitative models provide discipline and transparency, but managerial judgement remains essential because many decisive variables are difficult to measure precisely.

Learning Objectives

After studying this chapter, the reader should be able to:

1. Distinguish strategic, tactical, and operational location decisions.
2. Explain regional, national, and international location factors.
3. Apply factor-rating, centre-of-gravity, load–distance, break-even, and transportation models.
4. Construct a multi-criteria evaluation that combines financial and non-financial considerations.
5. Assess taxes, labour, infrastructure, political and climate risk, and sustainability.
6. Interpret location results critically and recognise model limitations.
7. Design a practical process for selecting and governing a facility network.

1. The Nature and Scope of Location Decisions

A facility location decision has three related dimensions:

• Geographical scale: region, country, city, industrial zone, or site.
• Structural scale: number, capacity, role, and ownership of facilities.
• Time scale: immediate placement, expansion, relocation, or long-term network evolution.

A single-facility decision asks where one plant or warehouse should be placed. A network decision asks which facilities should operate, what products or customers they should serve, how volumes should be allocated, and what transport links should connect them. Network design therefore includes location, capacity, sourcing, inventory positioning, transportation modes, and product flows.

Location decisions are typically classified as follows:

A useful distinction is between market-oriented and resource-oriented facilities. Market-oriented facilities are close to customers when delivery speed, freshness, or transport cost is important. Resource-oriented facilities are close to raw materials, energy, specialised labour, or suppliers. Some facilities are footloose, meaning that their location is relatively independent of a particular resource or market; digital services and certain assembly operations may fall into this category, although data connectivity and talent remain location constraints.

2. Location Factors Across Geographical Scales

Regional factors

At the regional level, the analyst compares broad economic and institutional environments. Important factors include market size and growth, labour availability and wage levels, industrial clusters, supplier depth, energy reliability, road and rail connectivity, ports, airports, taxes, public incentives, environmental standards, political stability, and exposure to natural hazards.

A regional comparison should examine both current conditions and trends. A region with low wages but declining population may face future labour shortages. A rapidly developing region may provide market growth but suffer from congestion and infrastructure bottlenecks.

National and international factors

International location choices introduce additional complexity:

• exchange-rate volatility and currency convertibility;
• tariffs, customs procedures, trade agreements, and rules of origin;
• corporate tax, withholding tax, value-added tax, and transfer-pricing rules;
• restrictions on foreign ownership or repatriation of profits;
• intellectual-property protection and data-sovereignty requirements;
• political stability, sanctions, corruption exposure, and regulatory predictability;
• labour law, working-time rules, collective bargaining, and occupational safety;
• cultural compatibility and managerial mobility;
• climate exposure, water availability, and carbon regulation.

Tax incentives must be evaluated as part of total economic value rather than treated as a guaranteed benefit. Incentives may be conditional on employment, investment, export volume, or operating duration. Their value may be reduced by clawback provisions, administrative delays, or changes in government policy.

Site-level factors

After a region and country have been selected, site-level analysis examines land cost, site preparation, zoning, building standards, access roads, utility capacity, drainage, security, expansion possibilities, and proximity to transport terminals. A site should be evaluated for its entire life-cycle cost, not only its purchase price. A cheap site requiring major road, power, water, or flood-protection investment may be economically inferior.

3. A Structured Location-Selection Process

A robust process normally includes the following stages:

1. Define the strategic purpose. Clarify products, customers, service promises, expected life, ownership model, and required capacity.
2. Specify decision criteria. Separate must-have constraints from desirable attributes.
3. Generate alternatives. Include existing facilities, brownfield sites, greenfield sites, outsourcing, and distributed configurations.
4. Screen infeasible alternatives. Eliminate options failing legal, capacity, environmental, or service requirements.
5. Collect comparable data. Establish common units, time periods, currency, and demand assumptions.
6. Apply quantitative models. Use cost, distance, capacity, and flow analyses.
7. Evaluate qualitative and uncertain factors. Include risk, labour relations, resilience, social licence, and sustainability.
8. Perform sensitivity and scenario analysis. Test demand, wages, exchange rates, transport prices, taxes, and disruption assumptions.
9. Conduct due diligence. Verify land title, permits, utilities, infrastructure, labour conditions, and incentive contracts.
10. Implement and review. Establish milestones, governance, performance measures, and contingency plans.

The distinction between a constraint and a criterion is important. A site lacking legally required environmental approval may be infeasible regardless of its score. By contrast, travel time or labour cost may be criteria that can be traded off against one another.

4. Quantitative Location Models

4.1 Factor-rating method

The factor-rating method converts diverse criteria into a weighted score. For alternative j, the weighted score is:

where w_i is the weight of criterion i, r_{ij} is the rating of alternative j on criterion i, and the weights sum to 1. Ratings may use a scale such as 1–10. The method assumes that criteria can be represented on a comparable scale and that weighted addition is an acceptable approximation of managerial preference.

Illustrative example. Three sites are rated on labour, transport, utilities, and risk. Weights are 0.30, 0.25, 0.20, and 0.25 respectively. Site A has ratings 8, 7, 9, and 6; Site B has 7, 9, 8, and 8.

Site A:

Site B:

Site B ranks higher. This conclusion is conditional on the weights, rating definitions, and completeness of the criteria. A sensitivity test should examine whether small changes in weights reverse the ranking.

A common weakness is false precision. A rating of 8 may not be exactly twice as desirable as a rating of 4. Analysts should document rating anchors and, where possible, use measured data rather than unsupported judgement.

4.2 Centre-of-gravity method

The centre-of-gravity method estimates a location that minimises weighted Euclidean distance under simplified assumptions. If customer or demand point i has coordinates (x_i, y_i) and weight w_i, the initial estimate is:

Weights may represent demand volume, shipment weight, or annual transport activity. Coordinates must use a consistent spatial system, and the method assumes that distance is reasonably represented by the coordinate geometry. It does not automatically account for roads, mountains, congestion, tariffs, capacity, or fixed facility costs.

Example. Three markets have coordinates and annual demand as follows:

Total demand is 600. Therefore:

The result is coordinate (6.5, 4.0), which is a screening estimate rather than an automatically feasible site. The nearest practical site must be identified using actual road distances, zoning, land availability, and service requirements.

4.3 Load–distance method

The load–distance method compares candidate sites using loads and distances:

where LD_j is the score for candidate site j, L_i is the load associated with point i, and d_{ij} is the distance between site j and point i. Lower scores are preferred when all loads and distance units are comparable.

If transport rates differ by lane, a cost-weighted version is more appropriate:

where c_{ij} is the cost per load-distance unit. The model is useful for comparing candidate warehouses, but it should not be used alone when fixed costs, capacity, multi-modal routes, or service constraints are substantial.

4.4 Break-even analysis

Break-even analysis compares fixed and variable costs. For alternative j:

where TC_j(Q) is total cost, F_j is annual fixed cost, v_j is variable cost per unit, and Q is annual volume. The intersection of two alternatives, A and B, occurs at:

provided the denominator is non-zero. The formula assumes linear variable cost, constant unit costs over the relevant range, and comparable output quality and service.

Example. Site A has fixed cost of $2.0 million and variable cost of $18 per unit. Site B has fixed cost of $3.2 million and variable cost of $14 per unit.

Below 300,000 units, A is cheaper; above 300,000 units, B is cheaper. This result excludes taxes, disruption cost, service penalties, and capacity limits unless they are incorporated into the cost functions.

4.5 Transportation and network-flow models

For a distribution network, the transportation model allocates supply from plants to markets at minimum variable transport cost. A basic formulation is:

subject to:

Here, x_{ij} is quantity shipped from plant i to market j, c_{ij} is unit shipping cost, S_i is available supply, and D_j is market demand. If total supply equals total demand, the inequalities may be written as equalities. Real network models often add facility-opening decisions, fixed costs, capacity, inventory, service-level, carbon, and risk constraints. Such mixed-integer models are powerful but require reliable data and computational expertise.

4.6 Multi-criteria decision analysis

A multi-criteria model combines financial, operational, social, environmental, and risk measures. A simple additive form is:

where u_{kj} is the normalised utility of alternative j on criterion k, and w_k is its weight. Benefit criteria are normally scaled so that larger values are better; cost or risk criteria must be reverse-scaled so that larger utility still means better performance.

One common min–max normalisation is:

for a benefit criterion. For a cost criterion:

These formulas require a non-zero range. If all alternatives have the same value, the criterion cannot discriminate and should be reviewed. Additive models assume preferential independence; a very poor score on one criterion may need to be treated as a veto or constraint rather than compensated by high scores elsewhere.

5. Taxes, Labour, Infrastructure, Risk, and Sustainability

Taxes and incentives

The relevant measure is often the after-tax, risk-adjusted net present value rather than nominal operating cost. A simplified project value is:

where I_0 is initial investment, CF_t is after-tax cash flow in period t, r is the discount rate, and t is the period. The model requires consistent currency, inflation treatment, tax assumptions, and timing conventions. It should include residual value, closure cost, remediation, working capital, and incentive clawbacks where material.

Labour

Labour analysis should include wage rates, productivity, absenteeism, turnover, recruitment time, training, industrial relations, health and safety, and future skill availability. Comparing hourly wages alone is misleading. A useful labour cost per good unit is:

The denominator should reflect actual productive output after downtime, quality losses, and learning effects.

Infrastructure and digital connectivity

Infrastructure includes roads, ports, rail, airports, power, water, waste treatment, telecommunications, cloud connectivity, and emergency services. Modern facilities may depend as much on data latency, cyber-security, and network redundancy as on physical transport. Infrastructure reliability should be assessed through historical outage data and recovery capability, not only rated capacity.

Risk and resilience

Location risk includes natural hazards, geopolitical disruption, supplier concentration, labour disputes, cyber incidents, disease outbreaks, and regulatory change. A simple expected-loss calculation is:

where p_s is the probability of scenario s and C_s is its consequence. Probabilities and consequences are uncertain, so scenario analysis and stress testing are often more informative than a single expected value. Resilience may justify duplicate suppliers, multiple ports, regional inventory, modular capacity, or geographically distributed facilities.

Sustainability and social licence

A location decision should consider greenhouse-gas emissions, water stress, land conversion, biodiversity, waste, local air quality, worker welfare, and community impacts. Transport emissions can be estimated as:

The emission factor must specify the transport mode, fuel basis, and reporting boundary. Sustainability is not merely a reputational issue. Carbon prices, disclosure rules, water restrictions, and community opposition can materially alter long-term cost and feasibility.

6. Network Design and Facility Roles

A network should be designed around facility roles rather than only geographic locations. Common roles include:

• Global production plant: serves multiple markets and exploits scale.
• Regional plant: balances proximity, responsiveness, and cost.
• Focused plant: specialises in a product, process, or customer segment.
• Distribution centre: consolidates inventory and outbound shipments.
• Cross-dock: transfers inbound goods to outbound vehicles with limited storage.
• postponement or configuration centre: delays final customisation until demand is clearer.

Centralisation usually lowers duplicated inventory and improves scale, but may increase distance, lead time, and disruption exposure. Decentralisation improves responsiveness and redundancy but may raise fixed cost and reduce utilisation. The appropriate design depends on demand variability, product value, service promises, postponement opportunities, and transport economics.

7. Illustrative Case: Regional Distribution Network Selection

This is a fictional illustrative case. Meridian Home Systems sells appliances to three national markets and expects annual demand of 600,000 units. Management must choose between a central distribution centre and a two-region network. The central option has annual fixed cost of $5.0 million and average variable distribution cost of $22 per unit. The two-region network has combined fixed cost of $7.4 million and variable cost of $17 per unit. At the expected volume, annual costs are:

The regional network appears cheaper by $0.6 million. Its break-even volume relative to the central option is:

Because expected demand exceeds 480,000 units, the regional design is financially attractive under the stated assumptions.

However, management then evaluates non-financial factors. The regional design requires two management teams, duplicate information systems, and greater coordination. It provides better delivery speed, reduces dependence on one site, and lowers average transport distance. A flood-risk assessment shows that one proposed regional site has significant exposure. The team therefore rejects that site and evaluates a nearby alternative with slightly higher rent but protected access and reliable power. A factor-rating exercise gives the protected alternative a lower land-cost score but a substantially higher resilience and infrastructure score. Sensitivity analysis shows that the regional network remains preferred unless demand falls below 480,000 units, transport cost decreases sharply, or the second site experiences major labour constraints.

The recommended approach is a two-region network with staged capacity: one facility is fully developed initially, while the second begins with modular capacity and expansion triggers tied to demand and service performance. Governance includes quarterly review of volume, utilisation, delivery time, carbon emissions, disruption incidents, and labour availability.

Learning points:

1. Fixed and variable costs can reverse the preferred alternative at different volumes.
2. A mathematically attractive site may fail a risk or infrastructure screen.
3. Network design should include capacity staging and option value.
4. Sensitivity analysis is essential when demand and transport costs are uncertain.
5. Sustainability and resilience can be operational and financial requirements, not optional additions.

8. Limitations, Governance, and Contemporary Practice

Location models can produce misleading results when data are inconsistent, criteria are double-counted, or assumptions are hidden. Common errors include using straight-line distance where road distance matters, treating tax incentives as permanent, ignoring ramp-up losses, omitting closure costs, and assuming average demand instead of testing variability.

Model governance should include documented data sources, version control, independent review, explicit assumptions, and approval thresholds. Decision makers should distinguish model output from recommendation: the model informs judgement but does not replace it.

Digital twins, geographic information systems, real-time freight data, and optimisation software increasingly support location analysis. Artificial intelligence can assist demand forecasting, geospatial screening, scenario generation, and anomaly detection. Nevertheless, AI outputs require validation because biased historical data, unstable relationships, and opaque recommendations can distort capital decisions. Sensitive labour, community, and geolocation data also require privacy and cybersecurity controls.

Contemporary practice increasingly emphasises regionalisation, nearshoring, dual sourcing, renewable-energy access, circular-economy infrastructure, and climate adaptation. The objective is shifting from minimum nominal cost to total value under uncertainty, incorporating service, resilience, sustainability, and strategic flexibility.

Chapter Summary

Facility location and network design determine the physical structure through which an organisation creates and delivers value. Decisions should be made across geographical, structural, and time scales, beginning with strategic purpose and ending with implementation governance. Regional and international analysis must consider markets, labour, infrastructure, taxes, regulation, risk, and sustainability.

Factor-rating and multi-criteria models support structured comparison of qualitative and quantitative criteria. Centre-of-gravity and load–distance methods provide spatial screening tools. Break-even analysis identifies volume thresholds between cost structures, while transportation and network-flow models allocate supply and demand subject to capacity and service constraints. Financial analysis should incorporate after-tax cash flows, incentives, investment, and life-cycle costs.

No single model is sufficient. Effective decisions combine quantitative analysis, scenario testing, due diligence, stakeholder considerations, resilience planning, and managerial judgement. The preferred network is usually the one that delivers the best risk-adjusted strategic value rather than the lowest immediate cost.

Key Terms

Facility location: Selection of the geographical position of an operating facility.

Network design: Configuration of facilities, flows, capacities, suppliers, markets, and transport links.

Market-oriented facility: Facility located near customers or demand centres.

Resource-oriented facility: Facility located near materials, energy, labour, or suppliers.

Factor rating: Weighted scoring method for comparing alternatives.

Centre of gravity: Weighted coordinate estimate based on demand or shipment weights.

Load–distance method: Comparison using loads multiplied by distances.

Break-even volume: Volume at which two alternatives have equal total cost.

Fixed cost: Cost that does not vary with short-run output volume.

Variable cost: Cost that changes with output or flow volume.

Transportation model: Optimisation model allocating supply to demand at minimum transport cost.

Multi-criteria decision analysis: Structured evaluation using several weighted criteria.

Resilience: Capacity to withstand, recover from, and adapt to disruption.

Scenario analysis: Examination of alternative future conditions and their consequences.

Tax incentive: Government benefit intended to encourage investment or employment.

Brownfield site: Previously developed site available for redevelopment or reuse.

Greenfield site: Previously undeveloped site selected for new construction.

Postponement: Delaying final product configuration until demand information improves.

Social licence: Community acceptance of an organisation’s presence and activities.

Total cost of ownership: Full life-cycle cost including acquisition, operation, risk, and closure.

Review Questions

Short Answer Questions

1. Distinguish between a facility location decision and a network design decision.
2. What is the difference between a constraint and a criterion in location analysis?
3. State the assumptions underlying basic break-even analysis.
4. Why can wage rate alone be an inadequate measure of labour cost?
5. Identify four risks that should be assessed in an international location decision.

Long Answer Questions

1. Explain the structured process for selecting a facility location from strategic definition through implementation.
2. Compare factor-rating, centre-of-gravity, load–distance, and transportation models, including their appropriate uses and limitations.
3. Discuss how taxes, incentives, exchange rates, and regulation affect international location decisions.
4. Analyse the trade-off between centralised and decentralised distribution networks.
5. Explain how sustainability and resilience can be incorporated into location and network design.

Analytical / Application Questions

1. Two sites have annual fixed and variable costs of ($2.0 million, $18 per unit) and ($3.2 million, $14 per unit). Determine the break-even volume and recommend a site for annual demand of 250,000 units and 450,000 units.
2. Using the centre-of-gravity method, calculate the coordinates for markets (2,8), (7,5), and (9,2) with demands 100, 200, and 300 respectively. Explain why the result may not be a feasible final site.
3. Design a multi-criteria evaluation for two competing international locations. Specify at least six criteria, their weights, measurement or rating method, and two sensitivity tests that could challenge the initial recommendation.

Multiple Choice Questions with Answers

Question 1

Which model estimates a location from weighted coordinates of demand points?

A. Break-even model

B. Centre-of-gravity method

C. Factor-rating method

D. Transportation model

Correct Answer: B. Centre-of-gravity method

Question 2

In a basic factor-rating model, the weights should normally sum to:

A. 10

B. 100

C. 1

D. The number of alternatives

Correct Answer: C. 1

Question 3

Two sites have fixed costs of $2,000,000 and $3,200,000 and variable costs of $18 and $14 per unit respectively. What is their break-even volume?

A. 120,000 units

B. 250,000 units

C. 300,000 units

D. 480,000 units

Correct Answer: C. 300,000 units

Question 4

In a transportation model, xᵢⱼ represents:

A. The fixed cost of plant i

B. The quantity shipped from i to j

C. The demand variance at market j

D. The distance between two sites

Correct Answer: B. The quantity shipped from i to j

Question 5

Which is a characteristic of a market-oriented facility?

A. It is located near a raw-material source

B. It is located near customers

C. It avoids all transport costs

D. It requires no demand information

Correct Answer: B. It is located near customers

Question 6

Which item is most appropriately treated as a feasibility constraint rather than merely a weighted preference?

A. Aesthetic quality of the site

B. Optional parking capacity

C. Legal environmental approval

D. Preferred office layout

Correct Answer: C. Legal environmental approval

Question 7

For a benefit criterion in min–max normalisation, a higher raw value generally receives:

A. A lower utility score

B. A higher utility score

C. A zero score automatically

D. An unchanged score

Correct Answer: B. A higher utility score

Question 8

What is a principal limitation of the centre-of-gravity method?

A. It cannot use demand weights

B. It automatically includes taxes

C. It may ignore roads, terrain, and site feasibility

D. It requires fixed costs for every facility

Correct Answer: C. It may ignore roads, terrain, and site feasibility

Question 9

Which statement best describes resilience in network design?

A. Minimising all inventory regardless of risk

B. The ability to withstand, recover from, and adapt to disruption

C. Selecting the site with the lowest land price

D. Using only one highly utilised supplier

Correct Answer: B. The ability to withstand, recover from, and adapt to disruption

Question 10

Why should tax incentives be treated cautiously in location analysis?

A. They are never legally available

B. They may be conditional or subject to clawback

C. They eliminate variable costs

D. They always increase labour productivity

Correct Answer: B. They may be conditional or subject to clawback