A distribution network is the system of storage locations, handling operations, and transportation links a company uses to move finished goods from the manufacturer to the final buyer (either directly or via intermediaries such as wholesalers and retailers). It covers physical sites (warehouses, fulfillment centers, sortation facilities), the transport modes that connect them (trucks, rail, air, vans, drones), and the information systems that coordinate flows and orders.
Key terms (brief)
– SKU: stock keeping unit, a unique identifier for a product or product variant.
– Fulfillment center: a warehouse focused on receiving customer orders, picking, packing, and shipping.
– Sortation center: a facility that groups parcels by delivery route or destination for efficient last-mile distribution.
– Hub-and-spoke: a design where a few central hubs serve many smaller regional facilities (“spokes”).
– Decentralized network: many regional or local facilities that operate more independently, closer to customers.
– Wholesaler/retailer: intermediaries that buy and resell products; wholesalers sell in bulk to businesses, retailers sell to consumers.
Why distribution networks matter
An efficient distribution network reduces delivery time, lowers transport and handling costs, and supports customer service targets. As online shopping and fast-delivery expectations grow, the network becomes a strategic asset: it affects margins, service levels, and capital spending (for example, on fulfillment centers or delivery fleets).
How companies organize distribution networks
– Centralized (single hub): lower facility overhead, simpler inventory management, but longer last-mile transit and potentially slower delivery.
– Hub-and-spoke: central regional hubs feed smaller local facilities; balances scale efficiencies with regional responsiveness.
– Decentralized (multiple regional centers): higher facility costs but faster delivery, lower last-mile transport distance, and greater resilience to local disruptions.
Companies often segment networks by product type (e.g., food vs. apparel) and by function (inbound receiving, storage, outbound picking, sortation, last-mile delivery).
Practical step-by-step guide to design or evaluate a distribution network
1. Map demand: collect customer locations, order volumes, seasonality, and service-level targets (e.g., next-day).
2. Segment products: identify fast-moving SKUs, temperature-sensitive items, bulky goods, and returns handling needs.
3. Select network topology: test centralized, hub-and-spoke, and decentralized options for your demand patterns.
4. Estimate costs: include facility fixed costs, inventory holding costs, transportation per-mile or per-shipment costs, labor, and IT/automation.
5. Model service metrics: estimate average transit time and variability, fill rates, and on-time delivery percentages.
6. Run trade-off scenarios: change number/location of centers, transport modes, and automation levels to find a cost/service balance.
7. Pilot and refine: deploy in a region, measure results, and scale changes iteratively.
Checklist — quick review before you decide
– Have you mapped customer density and delivery time targets?
– Did you segment SKUs by volume and handling needs?
– Have you compared fixed facility costs vs. variable transport costs?
– Do you have forecasts for peak seasons and returns?
– Are IT systems and data visibility specified (WMS, TMS, inventory accuracy)?
– Have you considered labor availability and local infrastructure quality?
– Have you modeled multiple scenarios and tested a pilot?
Worked numeric example — comparing centralized vs. decentralized
Assumptions
– Annual orders: 100,000
– Centralized option:
• Annual facility fixed cost: $2,000,000
• Average outbound transport cost per order: $10
– Decentralized option (4 regional centers):
• Combined fixed cost: $3,200,000
• Average outbound transport cost per order: $6
Calculate total annual cost and cost per order:
– Centralized total cost = fixed cost + transport cost = 2,000,000 + (100,000 × 10) = 3,000,000
• Cost per order = 3,000,000 / 100,000 = $30.00
– Decentralized total cost = 3,200,000 + (100,000 × 6) = 3,800,000
• Cost per order = 3,800,000 / 100,000 = $38.00
Interpretation
With these assumptions, the centralized design is cheaper per order. If transport savings in the decentralized model were larger or fixed costs lower
lowering the decentralized network’s combined fixed cost to below the centralized alternative (or increasing the per-order transport savings) would be required before the decentralized design becomes cost‑competitive on a total‑cost basis.
Inventory, service and lead‑time effects
– Inventory holding cost: Centralization typically increases average inventory (longer lead times and higher aggregate safety stock), which raises carrying costs. Carrying cost = average inventory × holding rate (percent) × unit cost.
– Service level and lead time: Decentralized facilities can shorten transit times and improve on‑time delivery and fill rates (the percentage of demand met from stock on hand). These service improvements have value beyond transport and facility costs.
– Risk and flexibility: Multiple facilities reduce exposure to single‑site disruption (supplier issues, weather, labor strikes) but increase operational complexity (management, IT, returns processing).
Worked sensitivity example — break‑even volume
We previously had:
– Centralized fixed Fc = $2,000,000, transport per order tc = $10
– Decentralized fixed Fd = $3,200,000, transport per order td = $6
To find the annual order volume Q at which total costs are equal:
Set Fc + tc·Q = Fd + td·Q
=> Q = (Fd − Fc) / (tc − td)
Plugging numbers:
Q = (3,200,000 − 2,000,000
) / (10 − 6) = 1,200,000 / 4 = 300,000 orders per year.
Interpretation
– If annual order volume Q > 300,000, the decentralized network (higher fixed cost but lower transport cost per order) yields lower total cost.
– If Q td and Fd > Fc for a positive Qbreak in this setup; otherwise interpret signs accordingly.)
Assumptions and caveats
– These calculations assume linearity: transport cost per order is constant and independent of volume (no quantity discounts), and fixed costs are truly fixed over the considered range.
– Inventory carrying costs, service-level benefits (e.g., higher sales from faster delivery), returns handling, taxes, and regulatory differences are omitted here but can materially change the outcome.
– Risk and resilience (value of geographic spread) are qualitative benefits that may justify higher cost even if the simple cost model favors centralization.
How to run a sensitivity check (step‑by‑step)
1. Gather inputs: realistic ranges for annual orders Q, fixed cost estimates for each layout (F), and transport cost per order (t). Include plausible best/worst values.
2. Compute Qbreak = (Fd − Fc) / (tc − td). Note how that value shifts when you vary F and t within your ranges.
3. Calculate TC for both networks at low, mid, and high demand scenarios (e.g., 25th, 50th, 75th percentiles).
4. Add secondary effects: estimate inventory carrying cost difference per order and estimate revenue upside from service improvement (if any). Convert those to per‑order or annual figures and include them in TC.
5. Create a simple table or chart of TC vs Q for both options to show the crossing point and margins.
6. Run stress cases: supply disruptions, transport price spikes, or order mix changes. Note where the preferred choice reverses.
Decision checklist (practical factors beyond the arithmetic)
– Demand volume and growth trajectory.
– Demand variability and seasonality.
– Value of faster delivery / higher fill rates (quantify if possible).
– Facility fixed costs and scalability (leases, land, labor).
– Transport cost structure (per order, per mile, LTL vs TL).
– Inventory carrying costs and safety stock required.
– Risk exposure (single‑site failure probability) and desired resilience.
– Complexity costs (IT, staff, reverse logistics).
– Regulatory/tax implications and local incentives.
Example sensitivity insight
– If transport per order difference shrinks
…shrinks, fixed facility costs and inventory carrying costs become the dominant drivers. That flips the economics toward fewer, larger sites because the marginal transport savings from splitting demand across more facilities no longer offset the extra facility and inventory costs.
Worked numerical break-even (step‑by‑step)
Use a simple total annual cost (TAC) model you can reproduce in a spreadsheet
TAC = Transport_cost + Facility_fixed_costs + Inventory_carrying_cost + Order_handling_cost
Where:
– Transport_cost = orders_per_year × transport_cost_per_order
– Facility_fixed_costs = number_of_facilities × fixed_cost_per_facility
– Inventory_carrying_cost = average_inventory_value × carrying_rate
(average_inventory_value ≈ annual_demand_value × safety_stock_days / 365)
– Order_handling_cost can be added per order if material.
Example scenario (numeric)
Assumptions:
– Annual orders = 100,000
– Value per order (avg goods) = $200 → annual_demand_value = 100,000 × $200 = $20,000,000
– Carrying rate = 25% (annual)
– Safety stock days: single hub = 7 days; two hubs = 5 days (lower lead‑time variability)
– Fixed_cost_per_facility: single hub = $1,000,000; each additional hub = $500,000
– Transport_cost_per_order: single hub = $8.00; two‑hub average = $5.00
– Order_handling cost ignored for simplicity
Compute inventory_carrying_cost:
– Single hub avg inventory = $20,000,000 × 7/365 ≈ $383,562 → carrying cost = $383,562 × 25% ≈ $95,891
– Two hubs avg inventory (split) = $20,000,000 × 5/365 ≈ $273,973 → carrying cost = $273,973 × 25% ≈ $68,493
Compute TAC:
– Single hub: Transport = 100,000 × $8 = $800,000; Facility = $1,000,000; Inventory = $95,891 → TAC ≈ $1,895,891
– Two hubs: Transport = 100,000 × $5 = $500,000; Facility = 2 × $500,000 = $1,000,000; Inventory = $68,493 → TAC ≈ $1,568,493
Result: In this example, two hubs are ≈$327K cheaper annually. If transport savings shrink (say the two‑hub transport cost drops from $5 to $7), then:
– Two hubs transport = $700,000 → TAC ≈ $1,768,493 → advantage falls to ≈$128K.
If the fixed cost per extra facility rises (e.g., to $700K), two hubs facility cost = $1.4M → TAC ≈ $1,968,493 → single hub wins.
Break‑even formula (for two vs one hub, ignoring handling costs)
Set TAC_one = TAC_two and solve for transport_cost_per_order_two (p2):
orders × p1 + F1 + IC1 = orders × p2 + 2×F2 + IC2
=> p2 = p1 + (F1 – 2F2 + IC1 – IC2) / orders
Plug values to get the transport per‑order threshold where configurations are equal.
Practical decision checklist (quick)
– Recompute TAC with realistic local labor and land costs.
– Run sensitivity ranges for transport_per_order, facility_fixed_cost, and safety_stock_days.
– Map geography: travel time vs distance; urban congestion raises per‑order transport cost.
– Consider service KPIs: lead time, fill rate, and acceptable customer experience.
– Evaluate operational complexity: more sites raise IT, staffing, and governance needs.
– Consider optional hybrid tactics: regional cross‑docks, pool distribution, or micro‑fulfillment.
Implementation steps (pilot approach)
1. Build a baseline TAC model with actual invoices, freight lanes, and SKU-level demand.
2. Run scenario runs: 1, 2, 3, N hubs; include seasonal high/low demand.
3. Select a lower‑risk pilot region or SKU cluster to test a new routing or a second facility.
4. Measure pilot KPIs for at least one full demand cycle (typically 3–6 months).
5. Scale iteratively, updating cost inputs and governance playbooks.
KPIs to monitor (frequency in parentheses)
– Total landed cost per order (monthly)
– Transport cost per order by lane (monthly)
– Fill rate / backorder rate (weekly)
– Average order lead time and variance (weekly)
– Inventory days of supply and turns (monthly)
– Facility utilization and labor hours per order (monthly)
– Disruption incidence and recovery time (event‑driven)
Common pitfalls and mitigations
– Over‑optimistic transport estimates: validate with carrier rate cards and historical lanes.
– Ignoring SKU heterogeneity: high‑value or fast movers may justify different placement rules.
– Underestimating complexity costs: add a burden for IT and management overhead in the model.
– Failing to stress‑test: simulate fuel spikes, port delays, or major customer changes to reveal fragility.
Final practical tip
Use the break‑even formula to compute a “transport savings needed per order” for each extra facility. That single number helps procurement and operations judge whether potential carrier negotiations or density improvements justify network changes.
Educational disclaimer
This is educational content for planning and teaching. It is not individualized investment or operational advice. Always validate models with your own data and consult relevant operational and financial professionals before implementing changes.
References
– Investopedia — Distribution Network:
– Council of Supply Chain Management Professionals (CSCMP) — Supply Chain Definitions and Resources:
– U.S. Bureau of Transportation Statistics — Commodity Flow and Freight Data:
– McKinsey & Company — Articles on supply‑chain network design:
– Harvard Business Review — Logistics and Distribution Strategy content