Logistics network optimization India means deciding how many hubs to operate, where to place them, how to structure lanes between them, and how to mix road, rail, air, and water modes. A well-designed network can cut landed cost by 15 to 25 percent versus a naive any-to-any model. The framework rests on five inputs: demand density by pin code, service-level promise, mode economics, multi-modal corridor access via PM GatiShakti, and reverse-flow volume.
Why network design is the biggest cost lever
Landed cost in Indian ecommerce decomposes roughly as 40 to 55 percent last-mile, 30 to 40 percent middle-mile, and 10 to 15 percent first-mile — within industry benchmark bands. Network design affects all three legs because every leg starts or ends at a node you chose. Wrong nodes drive wrong distances drive wrong cost.
The National Logistics Policy 2022 set the target of cutting national logistics cost from roughly 14 percent of GDP to 8 percent by 2030. That gap is closed shipper-by-shipper through better network design. A D2C brand running a poorly designed network can be carrying 20 percent extra cost relative to the same volume on an optimised one.
Our first-mile vs last-mile breakdown covers the leg-by-leg cost structure in depth. The principle here: logistics network optimization is about choosing the right nodes and lanes so each leg lands inside its benchmark band.
The five inputs you actually need
Real network design starts from five concrete inputs, not a software tool:
| Input | What it tells you | Where it surfaces |
|---|---|---|
| Demand density by pin code | Where customers actually are | Hub catchment design |
| Service-level promise by SKU class | How fast you must reach each customer | Hub count and lane structure |
| Mode economics (road/rail/air/coastal) | Cost per tonne-km per mode | Multi-modal mix |
| Corridor access (Gati Shakti, DFC) | Where multimodal links exist | Hub placement near corridors |
| Reverse-flow volume | How much returns add to throughput | Hub sizing and dedicated return lanes |
Skip any of these and the design will look right on paper but underperform on cost. The reverse-flow input is the most commonly missed.
Hub-and-spoke vs point-to-point vs hybrid
Three archetypes drive Indian network design. The decision logic, simply:
- Hub-and-spoke wins on consolidation density — small parcels and multi-shipper volume amortise hub cost efficiently.
- Point-to-point wins on transit time for high-AOV lanes — premium and intra-metro same-day cannot tolerate hub dwell.
- Hybrid is what most Indian carriers run — hub-and-spoke as default, direct lanes on top for high-volume pairs.
Practical decision rule: switch to direct when lane volume crosses approximately 5 to 7 tonnes per day on the route. Below that, consolidation through a hub is cheaper; above that, direct lanes start to win. For deeper coverage of the model itself see our hub-and-spoke logistics model explainer.
How to place hubs in India
The practical hub-placement heuristic that works for most D2C and B2B networks:
- Cluster pin codes by demand density. Top decile of pin codes by order volume = mandatory tier-1 node candidate. Next two deciles = tier-2 candidates. Bottom seven deciles = serve from nearest hub.
- Overlay PM GatiShakti and DFC access. Hubs near multimodal logistics parks or dedicated freight corridors carry a long-term cost advantage. The new MMLP infrastructure pipeline is the most important anchor for tier-2 hub placement decisions made today.
- Check fulfilment cost per node at expected throughput. A node only pays for itself above a volume threshold.
- Stress-test against alternative carrier hubs. Often the right answer is to partner with an existing carrier hub rather than build.
Most pan-India D2C operations need between four and seven hubs once they cross Rs 50 crore in revenue. A typical setup: two metro mega-hubs (Mumbai region + Delhi NCR), two regional hubs (Bengaluru + Kolkata or Hyderabad), and one to three tier-2 spokes based on category.
Multi-modal mix — when rail and coastal make sense
Most Indian logistics still moves by road. But for shippers with the right lanes, multi-modal mix can deliver step-change savings.
Rail freight is 30 to 60 percent cheaper than road on lanes over 800 km when volume permits container rakes. The constraint is first-mile and last-mile drayage to and from rail terminals — a CONCOR ICD or a private rail siding. The railway freight expansion coverage tracks the corridor build-out.
Coastal shipping is competitive for west coast to east coast moves — Mumbai to Chennai, Mundra to Vizag, and similar lanes. Sagarmala has been improving coastal connectivity. The trade-off is transit time and the need for inland connectivity at both ends.
Air freight retains a premium niche for high-value, time-definite, or low-density-by-weight shipments. UDAN cargo into tier-2 airports has expanded the addressable lanes.
Honest reality check: 90 percent plus of Indian ecommerce volume still moves by road. Multi-modal optimisation is a high-value lever for specific high-volume shippers, not a universal answer.
Tier-2 and tier-3 reach trade-off
Adding a tier-2 node cuts last-mile cost on its catchment by 20 to 30 percent but adds fixed opex. Break-even is typically at 800 to 1,200 orders per day from that catchment.
Below break-even, the right answer is to partner with regional last-mile players or use carrier-hub-attached fulfilment. The fixed cost of building before volume justifies it is a common D2C network design mistake.
The high-volume corridor lessons matter here: a brand routing significant volume on Mumbai to Delhi shipping and reverse should evaluate corridor-anchored hubs before any tier-2 build.
Reverse logistics in network design
Reverse flow in Indian ecommerce is typically:
- Fashion: 18 to 25 percent RTO of forward volume
- Electronics: 5 to 10 percent of forward volume
- FMCG and groceries: 1 to 3 percent of forward volume
Networks designed only for forward flow add 8 to 12 percent to total cost because returns chase the wrong nodes and consume the wrong capacity. Designing in reverse-collection PUDO points, dedicated return lanes, and consolidation hubs cuts the penalty.
A reverse-aware network design has three additions to the forward design: PUDO partnerships at the catchment, return-only outbound runs from regional hubs to mother warehouses, and consolidation of refurb/resale inventory at the right node. Most D2C networks under-invest here for two years before re-architecting.
The decision framework, in five steps
The end-to-end logistics network optimization framework:
- Map demand by pin code for the past 12 months by volume, AOV, and category.
- Classify SKUs by service tier — same-day, 24-hour, 48-hour, 5-day. Different tiers route differently.
- Shortlist nodes against PM GatiShakti / DFC access and MMLP pipeline. Reference our National Logistics Policy impact analysis for the corridor map.
- Cost each network configuration with mode mix — road default, rail/coastal where the lane and volume justify.
- Stress-test against reverse-flow volume and category-specific RTO.
Open-source linear programming (Python + PuLP) is good enough for most D2C networks. Commercial tools (Llamasoft, JDA, AnyLogistix) are typically overkill below Rs 500 crore revenue.
Frequently asked questions
What is logistics network optimization?
Logistics network optimization is the practice of deciding how many hubs to operate, where to place them, how to structure lanes between them, and how to mix road, rail, air, and coastal modes. A well-designed network can cut landed cost by 15 to 25 percent compared with a naive any-to-any structure for the same service-level promise.
How many hubs does a pan-India D2C operation need?
Most pan-India D2C operations need between four and seven hubs, depending on volume, SKU service tiers, and category. A typical setup is two metro mega-hubs, two regional hubs, and one to three tier-2 spokes. Below Rs 50 crore in annual revenue, partnering with carrier hubs is usually cheaper than building.
When does rail freight beat road in India?
Rail beats road on lanes over 800 km when volume supports container rakes, with cost savings of 30 to 60 percent per tonne. The constraint is first-mile and last-mile drayage to and from rail terminals. Concor and DFC corridors have made multi-modal mixes economically viable for high-volume D2C and ecommerce shippers.
How do you place a hub in India?
The practical method is to cluster pin codes by demand density, overlay PM GatiShakti and Dedicated Freight Corridor access, check the cost per order at the expected throughput, and break-even versus partnering with an existing carrier hub. Most networks need a hub at 800 to 1,200 daily orders from a catchment.
How does reverse logistics affect network design?
Reverse flow in Indian ecommerce is typically 18 to 25 percent of forward volume in fashion and 5 to 10 percent in electronics, so networks designed only for forward flow add 8 to 12 percent to total cost. Designing in PUDO collection points, dedicated return lanes, and reverse consolidation hubs is essential for any D2C network at scale.
Conclusion
Logistics network optimization is iterative. Revisit network design annually as volume, SKU mix, and category economics shift. The framework rests on five inputs and a clear-eyed view of where multi-modal alternatives pay back. For broader cluster context see our pillar on the Indian courier and logistics industry, or talk to the team at CourierBook for network design support.
Reference reading: NITI Aayog on goods mobility and PM GatiShakti for the corridor master plan.