Case: A Warehouse Network Redesign

A company can go from “next-day delivery in most cities” to “orders delayed, trucks half-empty, inventory stuck in the wrong place” without changing a single product. The culprit is often not the warehouse itself, but the network - where inventory sits, how it flows, and which customer promise each node is meant to serve.

The Big Picture

A warehouse network is the physical skeleton of service. Redesigning it is about matching where demand is, how fast customers expect delivery, and what it costs to hold and move inventory.

Core Explanation: How to Solve a Warehouse Network Redesign Case

Think of the case as a total-system redesign. A warehouse is only one node. The real system includes suppliers, factories, ports, mother warehouses, regional distribution centres, dark stores, stores, delivery partners and customers.

The clean consulting way to solve it is to move from demand to design, not from design to demand. In other words, do not start with “we need three warehouses.” Start with “who needs what, where, and how fast?”

Step 1: Map demand by geography, customer and SKU

Split demand into useful clusters: city, pin code, channel, customer type and SKU velocity. A redesign for B2B spare parts is not the same as a redesign for grocery, fashion or appliances.

Look for three patterns:

• Demand density: Are orders concentrated in a few metros or spread across many towns?
• SKU velocity: Which products are fast-moving, slow-moving, bulky, fragile or high-value?
• Service sensitivity: Which customers truly need same-day or next-day delivery, and which can wait?

Step 2: Define the service promise before the warehouse count

Service promise is the target experience: same-day, next-day, two-day, weekly replenishment, fill rate or emergency availability. The service target decides whether inventory must sit close to the customer or can be pooled centrally.

If the promise is loose, centralisation usually wins because inventory pooling reduces duplication. If the promise is tight, proximity matters, but costs and safety stock rise.

Step 3: Choose the network archetype

Most redesign cases ask you to compare three archetypes. The interviewer is checking whether you understand the trade-off, not whether you memorise jargon.

Step 4: Model total cost, not just transport cost

A warehouse redesign must include all major cost buckets. If you reduce outbound freight but double inventory and rent, the redesign has failed.

Step 5: Test feasibility and transition risk

A beautiful network on Excel can fail in reality. Check whether the company has the land, labour, transport partners, warehouse management system, supplier discipline and change-management capacity to execute the redesign.

This is where a consulting answer becomes practical. If you want the broader context of how such work moves from diagnosis to implementation, revise the engagement lifecycle from kickoff to handover.

Worked Example: First-Pass Warehouse Location

Suppose a company serves three demand zones. A simple centre-of-gravity calculation gives a first-pass location, weighted by order volume. It is not the final answer, but it is a useful starting point before checking roads, rent, service promises and constraints.

Weighted X = (20 x 200 + 10 x 300 + 90 x 500) / 1,000 = 52.

Weighted Y = (80 x 200 + 20 x 300 + 30 x 500) / 1,000 = 37.

So the first-pass location is near (52, 37). But an interview-ready answer adds the caveat: “I would now overlay road connectivity, real-estate cost, labour availability, supplier routes, service promises and risk.”

Definitions You Can Say in One Breath

BigBasket: Redesigning the Grocery Network for Speed

Situation. Online grocery has a difficult supply-chain problem: demand is frequent, baskets are mixed, freshness matters, and customers increasingly expect speed. A purely centralised warehouse can pool inventory efficiently, but it struggles when customers want rapid delivery across dense city pockets.

The move. BigBasket’s quick-commerce play required a more local fulfilment logic: smaller neighbourhood nodes for fast-moving items, larger upstream facilities for replenishment, and sharper assortment decisions by micro-market. The primary driver was proximity to demand. Supporting drivers included better demand forecasting, curated local assortment, rider routing, replenishment discipline and cold-chain handling where needed.

The lesson. The redesign was not simply “add more warehouses.” It was a network-role change. Large facilities remained useful for pooling and replenishment, while local nodes served urgent, high-frequency demand. That is the exact insight interviewers want: redesign the network by segment, not by instinct.

How AI Changes Warehouse Network Redesign

AI changes this topic by making the redesign more dynamic. Earlier, companies redesigned warehouse networks every few years. Now, demand shifts, delivery promises and inventory placement can be simulated and adjusted more frequently.

• Granular demand forecasting: Machine learning can forecast demand by pin code, SKU, day of week, weather pattern or local event. This improves which SKUs should sit in local nodes versus central pools.
• Scenario simulation: Optimisation models can compare “2 warehouses versus 5 warehouses,” different service promises, rent assumptions, fuel cost sensitivity and stock-out risk before physical changes are made.
• Smarter inventory placement: AI can recommend where to position inventory based on probability of sale, lead time, capacity constraints and substitution patterns.

Interview Relevance

Common Mistake