Distribution Network Design

Introduction and Context

Distribution network design is the strategic process of deciding how products should move from suppliers and production facilities to customers. It determines the number, location, ownership, capacity and operating role of facilities such as distribution centres, cross-docks, fulfilment centres, regional warehouses and retail outlets. It also determines transportation routes, inventory positioning, order-allocation rules and service commitments.

A distribution network is therefore more than a collection of warehouses. It is an integrated system linking customer demand, inventory, transport, information and facilities. A well-designed network balances two competing objectives:

• Efficiency: achieving low total logistics cost and high asset utilisation.
• Responsiveness: providing short lead times, high availability and reliable service.

Network design decisions are usually strategic. A facility may operate for ten to thirty years, while transport prices, demand patterns, technology and customer expectations may change much sooner. Consequently, a network should be designed using scenarios rather than a single forecast.

Learning Objectives

By the end of this chapter, you should be able to:

1. Explain direct shipment, hub-and-spoke, cross-docking, centralised and decentralised distribution structures.
2. Compare distribution alternatives using cost, service, capacity and risk criteria.
3. Apply centre-of-gravity and facility-location reasoning.
4. Calculate relevant network costs and interpret cost-to-serve.
5. Evaluate capacity and service-level implications.
6. Recommend a distribution network for a practical business situation.

1. The Strategic Logic of Distribution Networks

The design of a network begins with the customer proposition. Important questions include:

• How quickly must customers receive orders?
• What proportion of demand is predictable?
• Are products perishable, hazardous, bulky, valuable or customised?
• Is demand concentrated geographically or widely dispersed?
• Does the product require installation, returns handling or technical support?
• How much variety must be stocked close to customers?

A network must also reflect the product's functional characteristics. Stable, high-volume products can often be distributed through efficient consolidated flows. Fast-moving, uncertain or high-margin products may justify local inventory and rapid fulfilment. The same company may therefore use different structures for different product families or customer segments.

The principal trade-off is the relationship between facility count and total logistics cost. More facilities generally reduce outbound distance and improve response time, but increase facility overhead, inventory duplication, management complexity and inbound replenishment cost. Fewer facilities usually create economies of scale and inventory pooling, but may increase delivery distance and lead time.

Total network cost can be represented conceptually as:

where:

• TC = total relevant network cost per planning period;
• F = facility operating and fixed costs;
• H = handling and processing costs;
• I = inventory holding and pipeline inventory costs;
• T_in = inbound transportation cost;
• T_out = outbound transportation cost;
• P = purchasing or production-related distribution costs included in the decision;
• S = shortage, expediting and service-failure costs;
• R = returns, recycling and reverse-logistics costs.

The equation is an analytical framework rather than a universal accounting identity. Costs must be defined consistently, measured over the same period and included only when relevant to the decision. For example, sunk costs should normally be excluded from a short-term operating comparison but may be relevant in a long-term network redesign.

2. Major Distribution Structures

2.1 Direct Shipment

In a direct-shipment structure, products move from a supplier or manufacturing plant directly to the customer, often without passing through a distribution centre. Full truckload delivery, supplier drop-shipping and factory-to-retailer replenishment are examples.

Direct shipment is attractive when orders are large, customers are geographically concentrated, products are bulky or handling is expensive. It reduces intermediate handling and may shorten physical flow time. Its weaknesses include low vehicle utilisation when orders are small, limited consolidation and potentially high delivery cost for dispersed customers.

Direct shipment is particularly appropriate when:

• order quantities are large enough to fill transport economically;
• demand is relatively predictable;
• the customer base is limited;
• products require special handling;
• postponement or factory configuration is valuable.

2.2 Hub-and-Spoke

A hub-and-spoke network moves freight through one or more central hubs. Shipments from many origins are consolidated at a hub and then distributed to many destinations. Airlines, parcel carriers and express logistics providers frequently use this structure.

The hub creates consolidation economies and supports high network connectivity. However, it also creates dependency on the hub, additional handling and the possibility of congestion. A hub failure can affect a large part of the network, so resilience measures such as backup hubs, alternate carriers and emergency routing are important.

2.3 Cross-Docking

Cross-docking transfers products from inbound vehicles to outbound vehicles with little or no long-term storage. In a typical operation, inbound goods are received, identified, sorted by destination and moved directly to outbound staging lanes.

Cross-docking is not simply “fast warehousing.” It requires synchronised schedules, accurate advance shipment information, reliable suppliers, compatible packaging and sufficient shipment volume. It is well suited to retail replenishment, promotional products, fast-moving consumer goods and products with predictable demand.

Benefits include lower storage requirements, reduced handling, faster flow and lower inventory exposure. Risks include dependence on timing, limited tolerance for inbound disruption and the possibility that a missed inbound delivery causes an outbound service failure.

2.4 Centralised Distribution

A centralised network uses one major distribution centre or a small number of national facilities. Inventory is pooled, which can reduce safety stock because demand variability is aggregated. Centralisation also supports standardised processes, specialised automation and stronger managerial control.

Its disadvantages include longer outbound distances, exposure to regional disruptions and potentially poor service to remote customers. Centralisation works especially well when products have high value relative to weight, demand is uncertain but geographically substitutable, and customers accept moderate delivery lead times.

2.5 Decentralised Distribution

A decentralised network holds inventory in several regional or local facilities. It can offer shorter delivery times, reduced last-mile distance and better adaptation to local market conditions. It may also reduce the risk that one facility disruption affects the entire customer base.

The costs are additional buildings, labour, information systems, duplicated inventory and more complex replenishment. Decentralisation is often justified where customer density is high in multiple regions, delivery speed is commercially critical or transport costs are large relative to inventory cost.

2.6 Hybrid and Omnichannel Structures

Most modern networks are hybrid. A firm may use a national distribution centre for slow-moving products, regional facilities for fast-moving lines, direct shipment for large business customers and stores as fulfilment points for online orders. Such designs require clear inventory ownership, allocation rules and visibility across channels.

3. Facility-Location Analysis

Facility-location analysis identifies suitable sites and assigns customer demand to facilities. It combines quantitative analysis with qualitative judgement. A mathematically attractive site may be unacceptable because of labour shortages, land-use restrictions, poor road access, security risks or environmental constraints.

3.1 Centre-of-Gravity Method

The centre-of-gravity method provides an initial geographical estimate for a single facility serving known demand points. In its basic form:

where:

• xᵢ, yᵢ = coordinates of customer or demand point i;
• wᵢ = weight assigned to point i, commonly annual demand, shipment weight or transport volume;
• x^*, y^* = calculated location coordinates.

The method assumes that transport cost is approximately proportional to volume and distance, routes are reasonably represented by straight-line or comparable distances, and there are no major geographic or regulatory constraints. It does not prove that the calculated point is feasible or optimal. It should be followed by road-network analysis, land and labour assessment, tax review, risk analysis and cost modelling.

Worked Example: Centre of Gravity

A company serves three markets:

Total demand is 1,000 tonnes. Therefore:

The initial estimate is therefore coordinate (26, 38). This point is only a screening result. If it lies in a protected area or lacks suitable road access, the firm must evaluate feasible sites nearby.

3.2 Factor-Rating Method

The factor-rating method compares candidate sites using weighted criteria. Let sⱼ be the score of site j for criterion i, and wᵢ be the criterion weight, where the weights sum to 1. The total score is:

Scores may use a scale such as 1 to 10. The method is useful for incorporating qualitative considerations, but results depend on judgement and should be tested through sensitivity analysis.

Typical criteria include transport access, labour availability, land cost, utility reliability, customer proximity, tax conditions, security, environmental exposure and expansion potential.

3.3 Network Optimisation Models

For larger decisions, firms may use a transportation or mixed-integer optimisation model. A simplified facility-location model minimises fixed facility costs and variable allocation costs:

subject to:

Here, Fⱼ is the fixed cost of opening facility j, zⱼ indicates whether it is opened, cᵢⱼ is the variable cost of serving customer i from facility j, qᵢⱼ is the quantity allocated, dᵢ is customer demand and Kⱼ is facility capacity. This basic model assumes known demand, linear variable cost and divisible flows. Real applications may require vehicle constraints, multiple products, service-time restrictions, inventory decisions and uncertainty modelling.

4. Capacity and Inventory Positioning

Capacity is the maximum usable flow or storage that a facility or transport resource can support during a specified period. It may be measured in pallet positions, cubic metres, cases per day, orders per hour, dock doors or labour hours.

A facility should not normally be designed at 100% average utilisation. Variability, maintenance, absenteeism, seasonal peaks and arrival bunching require operational headroom. A simple utilisation measure is:

where U is utilisation, D is demand or throughput during the period and K is available capacity in the same units and period. For example, 8,000 cases per day processed against practical capacity of 10,000 cases per day gives U = 0.80, or 80%.

Design capacity is the theoretical maximum. Effective capacity reflects staffing, maintenance, product mix and operating restrictions. Planning should use effective capacity and test peak demand, not only annual averages.

Inventory positioning is closely related to network design. Centralisation pools uncertainty, while decentralisation places inventory closer to demand. The risk-pooling effect is strongest when demand across locations is not perfectly correlated. However, transport responsiveness, replenishment frequency and product criticality may justify local stock even when pooling would reduce inventory.

5. Service-Level Analysis

A service level expresses the probability or proportion of demand served under a defined condition. The definition must be explicit. Common measures include:

• Cycle service level: probability of not experiencing a stockout during a replenishment cycle.
• Fill rate: proportion of demand quantity supplied immediately from available stock.
• On-time delivery: proportion of deliveries arriving within the agreed time window.
• Order completeness: proportion of orders delivered without missing lines.
• Perfect-order rate: orders delivered on time, complete, damage-free and with correct documentation.

These measures are not interchangeable. A company may achieve a high on-time rate but a low fill rate if deliveries are punctual but incomplete.

Safety stock can be approximated for a single item when demand during lead time is normally distributed:

where SS is safety stock in units, z is the standard-normal service factor and σ_LT is the standard deviation of demand during lead time. If daily demand has standard deviation σ_d, lead time is constant at L days, and daily demands are independent, then:

These formulas assume a stable demand distribution, independent daily demand and reliable measurement of lead time. If lead time is variable, demand and lead time uncertainty should be modelled together rather than using the constant-lead-time formula.

Higher service targets increase safety stock disproportionately. A move from 90% to 98% service may require a substantial inventory increase, depending on demand variability. The appropriate target should be based on customer value, shortage consequences and cost-to-serve rather than an arbitrary universal percentage.

6. Cost-to-Serve Analysis

Cost-to-serve (CTS) assigns the resources consumed by a customer, channel, product or order profile. It improves decisions that revenue and gross margin alone can distort. A customer with high sales may be unprofitable if it requires small frequent deliveries, special packaging, urgent transport, extensive returns or manual administration.

A simplified contribution measure is:

Distribution cost may include order processing, picking, packing, handling, storage, inbound and outbound transport, delivery failure, returns and customer-specific requirements. Activity-based costing can improve accuracy by assigning cost drivers such as order lines, pallets, delivery stops, kilometres, storage days and returns.

CTS should not be used mechanically to reject low-volume customers. Some customers may be strategically important, generate future growth or provide network density. Managers should distinguish between avoidable cost, allocated overhead and strategic value.

7. A Structured Network-Design Process

A professional design study normally follows these stages:

1. Define scope and objectives. Specify products, markets, planning horizon, service promises and cost categories.
2. Collect and validate data. Gather demand by customer and time period, product dimensions, transport rates, facility costs, lead times and capacity.
3. Segment demand. Separate customers and products by service need, volume, value, uncertainty and handling requirements.
4. Map the current network. Document facilities, flows, inventory points, transport lanes, constraints and performance.
5. Develop alternatives. Include direct, centralised, decentralised, hub, cross-dock and hybrid structures.
6. Model costs and service. Calculate facility, transport, inventory, handling, shortage and implementation costs.
7. Test scenarios. Examine growth, seasonality, fuel prices, supplier disruption, demand shifts, labour changes and service targets.
8. Assess qualitative factors. Consider resilience, regulatory compliance, workforce, sustainability, security and strategic flexibility.
9. Select and implement. Use staged implementation, governance, milestones and clearly assigned accountability.
10. Monitor and redesign. Networks require periodic review because demand and operating conditions change.

Illustrative Case: Meridian Home Appliances

Background

Meridian Home Appliances sells refrigerators, washing machines and small appliances to national retailers and online customers. Its existing network consists of one central warehouse near the manufacturing plant. The company promises two-day delivery to major cities but frequently uses premium transport for distant orders.

Problem

Sales have grown in three regional markets. The central warehouse has adequate annual capacity but experiences peak-season congestion. Large appliances are expensive to transport over long distances, while small appliances can be shipped economically through parcel carriers. Management is considering a second regional facility or a cross-dock arrangement.

Analysis

The company first segmented products. Large appliances were assigned to a regional stock-holding model because their customer service requirement and transport cost justified proximity. Small appliances remained centralised because inventory pooling and parcel consolidation were more valuable. Demand-weighted coordinates were calculated for regional markets, then adjusted to feasible sites near motorway junctions.

The analysis included:

• fixed rent, labour and equipment costs;
• inbound replenishment from the plant;
• outbound delivery cost by product family;
• inventory duplication and safety stock;
• two-day service achievement;
• peak capacity and dock requirements;
• disruption exposure;
• reverse logistics for damaged appliances.

A centre-of-gravity result identified an area near the largest regional market, but a factor-rating assessment preferred a nearby site with better labour availability and lower flood exposure. The firm also found that a full regional warehouse would create excessive small-appliance inventory. It therefore designed a hybrid solution.

Approach and Solution

The selected design used:

• the existing central warehouse for small appliances and slow-moving large appliances;
• a regional facility holding fast-moving large appliances;
• scheduled milk-run replenishment from the plant;
• cross-docking for selected retailer promotions;
• direct shipment for full truckload retailer orders;
• shared inventory visibility and order-allocation rules.

The company set different service targets by segment rather than applying a single target to every product. Capacity was planned using peak-week volume and a practical utilisation ceiling below theoretical maximum.

Outcome

The hybrid design reduced long-distance deliveries for bulky products, improved regional delivery reliability and avoided unnecessary duplication of the entire product range. Its success depended on accurate inventory records, disciplined replenishment and clear rules for deciding whether an order was served from the central facility, regional facility or directly from the plant.

Learning Points

The case illustrates that network design is not a choice between “one warehouse” and “many warehouses.” Product segmentation, cost-to-serve, service differentiation and capacity analysis can produce a more economical hybrid structure. It also demonstrates why a mathematically attractive location must be tested against labour, environmental and operational realities.

8. Technology, Sustainability and Governance

Modern network design uses geographic information systems, transport-management systems, warehouse-management systems, digital twins and optimisation software. Machine-learning tools can improve demand forecasts, identify delivery patterns and test scenarios. However, analytical outputs remain dependent on data quality, assumptions and governance. Managers should document model versions, data sources, objective functions and decision rules.

Artificial intelligence can support facility-location analysis, but it should not make unreviewed decisions involving employment, safety, regulatory compliance or customer access. Human review is necessary where data are incomplete or where an apparently efficient solution transfers cost or risk to vulnerable communities.

Sustainability should be assessed across the network rather than by considering vehicle distance alone. Relevant effects include building energy, refrigeration, packaging, empty running, congestion, land use, emissions and product returns. A common transport-emissions approximation is:

For transport, activity may be tonne-kilometres and the emission factor may be kilograms of CO₂-equivalent per tonne-kilometre. The boundaries, vehicle type, load factor and fuel assumptions must be stated. Consolidation can reduce emissions, but a decentralised facility may reduce last-mile distance; the net result requires measurement.

Governance issues include hazardous-material regulations, occupational safety, data protection, customs compliance, accessibility, fair labour practices and business-continuity obligations. Facility decisions can affect local traffic, noise and employment, so stakeholder consultation and environmental assessment may be appropriate.

9. Challenges and Limitations

Distribution network models are powerful but imperfect. Common limitations include:

• forecasts may be wrong or biased;
• transport rates may change with fuel and labour conditions;
• average demand hides peak congestion;
• linear cost assumptions may ignore volume discounts and capacity thresholds;
• service quality may be difficult to quantify;
• site scores may reflect subjective judgement;
• models may omit disruption and recovery costs;
• local optimisation may damage total supply-chain performance.

Sensitivity analysis is therefore essential. Managers should ask which assumptions would change the recommendation. If a solution is preferred only under one optimistic forecast, it may be less robust than a slightly more expensive design with greater flexibility.

Best-Practice Principles

1. Start with customer service requirements, not warehouse preferences.
2. Segment products and customers before choosing a structure.
3. Use total cost rather than transport cost alone.
4. Distinguish average capacity from peak practical capacity.
5. Treat location models as screening tools unless all constraints are represented.
6. Evaluate resilience, sustainability and regulatory obligations explicitly.
7. Use consistent cost and service definitions.
8. Test multiple scenarios and conduct sensitivity analysis.
9. Design information visibility and governance alongside physical facilities.
10. Review the network periodically as demand, technology and risk change.

Chapter Summary

Distribution network design determines how facilities, inventory, transport and information are configured to meet customer demand. Direct shipment minimises intermediate handling, while hub-and-spoke structures provide consolidation and broad connectivity. Cross-docking accelerates flow but requires reliable timing and accurate information. Centralised networks pool inventory and exploit scale; decentralised networks improve proximity and responsiveness but increase fixed and inventory costs. Hybrid designs often provide the best fit across different products and customer segments.

Facility decisions can be supported by centre-of-gravity, factor-rating and optimisation models. Centre-of-gravity calculations provide an initial demand-weighted location, whereas optimisation models can incorporate fixed costs, variable flows and capacity. These tools must be supplemented by feasibility, risk, sustainability and stakeholder assessment.

Capacity analysis must use practical peak capability rather than only annual averages. Service-level analysis requires precise definitions, and inventory decisions should recognise the relationship between uncertainty, pooling and safety stock. Cost-to-serve reveals the real resource consumption of customers and channels. Finally, effective network design is iterative: it combines quantitative modelling, managerial judgement, technology, governance and continuous review.

Key Terms

Distribution network: The facilities, transport links, inventory points and information flows used to move products to customers.

Direct shipment: Movement from origin to customer without an intermediate distribution facility.

Hub-and-spoke: A structure in which flows are consolidated through central hubs.

Cross-docking: Transfer from inbound to outbound transport with minimal storage.

Centralised distribution: Concentration of inventory and fulfilment in one or a few facilities.

Decentralised distribution: Placement of inventory in several regional or local facilities.

Hybrid network: A network combining different structures for different products or customers.

Centre of gravity: A demand-weighted geographical estimate for a facility location.

Factor rating: A weighted scoring method for comparing candidate sites.

Facility capacity: The usable processing, storage or throughput limit of a facility.

Utilisation: Demand or throughput expressed as a proportion of available capacity.

Inventory pooling: Combining demand across locations to reduce the effect of variability.

Service level: A defined measure of availability, delivery performance or order fulfilment.

Fill rate: The proportion of demand supplied immediately from available inventory.

Safety stock: Inventory held to protect against demand or lead-time uncertainty.

Cost-to-serve: The cost of activities and resources required to serve a customer, product or channel.

Network resilience: The ability to absorb disruption, continue service and recover performance.

Reverse logistics: The movement of products, packaging or materials from customers back through the supply chain.

Sensitivity analysis: Testing how a decision changes when assumptions or inputs change.

Review Questions

Short Answer Questions

1. What is the principal trade-off between centralised and decentralised distribution?
2. Define cross-docking and identify two conditions required for its success.
3. What does the centre-of-gravity method calculate?
4. Distinguish between fill rate and on-time delivery.
5. Why should practical peak capacity be used in network design?

Long Answer Questions

1. Compare direct shipment, hub-and-spoke and cross-docking structures, including their advantages, limitations and suitable applications.
2. Explain how inventory pooling influences the choice between centralised and decentralised distribution.
3. Describe a complete distribution network-design process from problem definition to implementation.
4. Discuss the role and limitations of facility-location models in strategic network decisions.
5. Explain how cost-to-serve analysis can change managerial conclusions about customer and channel profitability.

Analytical / Application Questions

1. Using the centre-of-gravity method, calculate the location coordinates for a facility serving three markets with coordinates (10, 20), (50, 30) and (30, 70), and annual demands of 400, 300 and 300 tonnes respectively. Interpret the result and identify two feasibility checks.
2. A company is considering one national warehouse or three regional warehouses. Develop a decision framework covering facility cost, transport, inventory, capacity, service, resilience and sustainability. State what additional data would be required.
3. Design a hybrid distribution network for an omnichannel retailer selling bulky slow-moving products and small fast-moving products. Justify the structure, inventory locations, service metrics and governance controls.

Multiple Choice Questions with Answers

Question 1

Which distribution structure moves products from inbound vehicles to outbound vehicles with minimal storage?

A. Cross-docking

B. Decentralised warehousing

C. Direct shipment

D. Reverse logistics

Correct Answer: A. Cross-docking

Question 2

In the centre-of-gravity method, the weighting variable is commonly based on which factor?

A. Facility age only

B. Annual demand or transport volume

C. Number of employees at headquarters

D. Product selling price only

Correct Answer: B. Annual demand or transport volume

Question 3

A major advantage of a centralised distribution network is:

A. Guaranteed shorter delivery distance to every customer

B. Elimination of all transport costs

C. Inventory pooling and economies of scale

D. No dependence on information systems

Correct Answer: C. Inventory pooling and economies of scale

Question 4

If a facility processes 8,000 cases per day and has practical capacity of 10,000 cases per day, its utilisation is:

A. 20%

B. 50%

C. 80%

D. 125%

Correct Answer: C. 80%

Question 5

Which measure represents the proportion of demand supplied immediately from available inventory?

A. Fill rate

B. Facility utilisation

C. On-time delivery

D. Inventory turnover

Correct Answer: A. Fill rate

Question 6

In a basic facility-location model, the binary variable indicating whether a facility is opened is commonly represented by:

A. qᵢⱼ

B. dᵢ

C. Kⱼ

D. zⱼ

Correct Answer: D. zⱼ

Question 7

Which situation most strongly supports direct shipment?

A. Many small orders dispersed across remote customers

B. Large orders that economically fill transport vehicles

C. Highly variable demand requiring pooled inventory

D. A requirement for extensive order consolidation

Correct Answer: B. Large orders that economically fill transport vehicles

Question 8

If daily demand standard deviation is σ_d and lead time is constant at L independent days, the standard deviation of lead-time demand is:

A. σ_d ÷ L

B. σ_dL

C. σ_d√L

D. σ_d + L

Correct Answer: C. σ_d√L

Question 9

Why can decentralised distribution improve customer service?

A. It always reduces total inventory

B. It places inventory closer to demand locations

C. It eliminates facility fixed costs

D. It removes the need for replenishment

Correct Answer: B. It places inventory closer to demand locations

Question 10

A key limitation of a centre-of-gravity result is that it:

A. Cannot use demand as a weighting factor

B. Always identifies the legally approved site

C. Ignores all transportation considerations

D. May identify an impractical point requiring further feasibility analysis

Correct Answer: D. May identify an impractical point requiring further feasibility analysis