DC Network Design

DC network design determines the number, location, size, and role of distribution and fulfillment nodes in a supply chain. It is the highest-leverage structural decision in logistics — a 10% reduction in average outbound zone is worth more over 10 years than most automation investments.

Total Landed Cost (TLC) Model

Network optimization minimizes total landed cost across four layers:

Layer	Typical % of TLC	Key Driver
Inbound freight	15–25%	Supplier locations, mode mix, lead time
Outbound freight	35–50%	Customer density, service level, zone distribution
DC operating cost	20–35%	Labor market, automation level, lease rates
Inventory carrying cost	10–20%	Safety stock multiplication across nodes (adding a node increases system inventory 20–30%)

Adding a DC node reduces outbound freight but increases inventory (safety stock scales with √N nodes) and fixed operating cost. The optimization finds the cost-minimizing node count.

Service Level Radii

Transit time commitments drive node placement:

Service Level	Radius from DC	Population coverage (US)
Same day	25–50 miles	Metro only
Next day	125–175 miles	~60% with 6–8 nodes
2-day	350–450 miles	~95% with 3–4 nodes
3-day (ground)	600–800 miles	~98% with 2 nodes

The continental US can be covered at 2-day ground with 3–4 strategically located DCs (typically: mid-Atlantic, Midwest, Southwest, Pacific Northwest or similar).

Network Design Triggers

Trigger	Signal
Service level gaps	Customer complaints about transit time; competitor advantage
Cost per order increasing	Rising outbound freight as % of revenue
M&A / footprint change	Acquired network overlaps or gaps
E-commerce growth	Shift from B2B pallet to B2C parcel changes optimal node count
Labor market shift	Wage inflation in current markets exceeds relocation cost
Lease expiration	Renewal cost triggers strategic review

Analytical Methods

Center-of-gravity (COG) analysis: Weighted average of customer locations by volume. Quick first-pass to identify optimal single-node location. Excel solvable; ignores costs, capacity, and multi-node interactions.

Gravity model (multi-node): Iterates COG across N nodes simultaneously. Most network design tools (LLamasoft/Coupa Supply Chain Guru, IBM Sterling, ILOG, Llamasoft) use this as the foundation, layered with TLC cost modeling and service constraints.

Mixed-integer programming: Formal optimization formulation. Used for complex networks with discrete facility choices, capacity constraints, and multi-echelon inventory.

Make vs. Buy

Factor	Favor Owned	Favor 3PL
Volume stability	High / predictable	Variable / seasonal
Operational control	Critical (hazmat, temp, specialized)	Standard commodities
Capital availability	CapEx available	CapEx constrained
Speed to market	Acceptable lead time	Need rapid deployment
Scale in region	Sufficient to justify fixed cost	Insufficient for standalone DC

See CapEx vs OpEx in Logistics for the lease vs. own framing.

Greenfield vs. Retrofit

Greenfield (new-build) allows clear-height, column spacing, dock ratio, and automation optimization from scratch. Retrofit to an existing building constrains design but avoids development timeline (12–18 months vs. 6 months for retrofit). Most network redesigns use existing buildings where possible; greenfield is reserved for strategic anchor nodes where no suitable existing facility exists.

Link to Site Selection

Network design determines where a node should be. Site selection then identifies the specific building or land parcel within that geography. See Manufacturing Plant Site Selection for the 4-phase site selection methodology (strategic filters → TLC → weighted scoring → due diligence).