Supply Chain Design, Network & Inventory Decisions
One retailer opens hundreds of stores and still keeps shelves full; another expands into the same cities and suddenly drowns in stockouts, excess inventory, and emergency freight. The difference is rarely “better execution” alone - it is the invisible architecture of supply chain design.
- Supply chain design decides the structure of the chain - facilities, capacity, inventory, transport, information and partners.
- Network decisions answer: how many plants, warehouses or dark stores, where, with what capacity, serving which demand zones.
- Inventory decisions answer: what to stock, where to stock it, how much to reorder, and what service level to promise.
- The core trade-off is cost efficiency vs responsiveness: centralisation lowers inventory cost; decentralisation improves speed and local availability.
- Use a structured answer: demand profile - service promise - network options - inventory policy - cost and risk trade-offs - KPIs.
- Track hard measures: service level, fill rate, OTIF, inventory turns, days inventory, logistics cost as % of sales.
- The common trap: recommending “more warehouses” or “more inventory” without linking it to demand variability, service promise and total cost.
Big Picture - The Architecture Before the Trucks Move
A supply chain is not designed one warehouse at a time. It is designed by matching the customer promise with the physical network and the inventory logic needed to deliver that promise profitably.
Core Explanation - The Three Decisions That Matter
Think of supply chain design as three connected decisions: where the chain is located, how goods flow, and where inventory waits. If one is wrong, the others compensate expensively.
1. Design Decisions - The Strategic Choices
These are long-term choices that are expensive to reverse. They shape the cost base and service capability for years.
If you are solving a network redesign case, combine this with a warehouse network redesign case framework so that you do not jump from symptoms to recommendations too quickly.
2. Network Decisions - Where the Chain Lives
Network design asks: how many nodes should exist, where should they be, and which customers should each node serve? A centralised network is usually cheaper; a decentralised network is usually faster.
A strong answer does not say one side is “better”. It says: centralise what benefits from pooling, decentralise what benefits from speed.
3. Inventory Decisions - Where Cash Becomes Service
Inventory is both a buffer and a cost. It protects the customer promise against uncertain demand and supply, but it also locks working capital, space and handling effort.
The main inventory types are:
- Cycle inventory: stock created because companies order or produce in batches.
- Safety inventory: buffer stock held against demand or lead-time uncertainty.
- Seasonal inventory: stock built ahead of predictable spikes such as festivals or sale events.
- Pipeline inventory: stock already in transit between two nodes.
- Decoupling inventory: stock placed between process stages so one delay does not stop the entire chain.
The Supply Chain Fit Matrix
A good network design matches demand uncertainty and supply uncertainty. Predictable demand with stable supply needs efficiency. Uncertain demand with unreliable supply needs agility and buffers.
For example, basic staples can run through an efficient replenishment system. Trend-led apparel, premium electronics or quick-commerce essentials need a more responsive network because the penalty of being late or stocked out is high.
Key Metrics to Track - Cost, Service and Inventory Health
Supply chain design must be judged with numbers. A network that reduces freight but destroys availability is not a win; inventory reduction that creates stockouts is just cost cutting disguised as efficiency.
If the case involves whether a warehouse is overloaded, revise capacity, throughput and utilisation analysis before you recommend adding more facilities.
Worked Example - EOQ and Safety Stock in 3 Minutes
Suppose a retailer sells a hypothetical SKU with annual demand of 12,000 units. Ordering cost is ₹500 per order and annual holding cost is ₹20 per unit.
Economic Order Quantity:
EOQ = √(2DS / H)
- D = annual demand = 12,000 units
- S = ordering cost = ₹500 per order
- H = holding cost = ₹20 per unit per year
EOQ = √(2 × 12,000 × 500 / 20) = √600,000 ≈ 775 units.
Interpretation: if assumptions are stable, ordering about 775 units balances ordering cost and holding cost. In a real case, you would adjust for minimum order quantities, supplier reliability, shelf life, storage constraints and service level.
Simple safety stock: if demand during lead time has a standard deviation of 100 units and the company wants roughly 95% cycle service, use z ≈ 1.65.
Safety stock = z × σLT = 1.65 × 100 = 165 units.
So the average planned stock logic becomes: cycle stock around half the order quantity, plus safety stock for uncertainty.
Definitions You Can Say in One Breath
The Council of Supply Chain Management Professionals defines supply chain management as “the planning and management of all activities involved in sourcing and procurement, conversion, and all logistics management activities” (CSCMP Supply Chain Management Definitions).
- Supply chain design: The structural choices that determine facilities, flows, capacity, inventory and partners across the supply chain.
- Network design: The decision on number, location, role and capacity of supply chain nodes.
- Inventory policy: Rules for what to stock, where to stock, when to reorder and how much to order.
- Service level: The probability or frequency with which demand is fulfilled without a stockout.
- Cost to serve: The total cost of fulfilling a customer, channel, SKU or region.
Case Study - DMart: Low-Cost Retail Built on Network Discipline
DMart shows how a retailer’s customer promise, store network and inventory discipline must reinforce one another to make everyday low pricing work.
Situation: Indian grocery and household retail is brutally price-sensitive. Customers expect availability, but the retailer cannot afford bloated inventory, frequent markdowns or expensive emergency replenishment.
The move: DMart’s operating logic has been to keep the promise simple: value pricing, a focused assortment, disciplined store economics and strong replenishment control. The supply chain implication is clear: stores must be supplied reliably, SKUs must move fast, and the network must expand in a way that the back end can support.
Primary driver: The biggest driver is alignment between business model and supply chain design. A low-price retailer cannot run a scattered, high-complexity network and still protect margins.
Supporting drivers: The model is supported by cluster-based expansion, tight SKU discipline, supplier negotiation strength, store-level demand visibility, and inventory turns that convert stock into cash quickly.
Lesson: DMart is not just a retail pricing story. It is a supply chain design story where network density, SKU discipline, supplier economics and inventory velocity reinforce the same strategic promise.
How AI Changes Supply Chain Design, Network & Inventory Decisions
AI does not remove the classic trade-offs. It makes them more visible, faster to simulate and easier to monitor.
- Demand sensing: Machine learning models can combine sales history, promotions, weather, local events and search signals to improve short-term forecasts at SKU-location level.
- Inventory optimisation: AI can recommend different safety stock levels for different SKUs instead of applying one blanket rule across the network.
- Network digital twins: Companies can simulate facility closures, demand shifts, transport delays or new warehouse locations before committing capital.
- Exception management: AI copilots can flag likely stockouts, delayed inbound shipments or abnormal demand spikes and suggest actions.
Student workflow: Use NotebookLM or Claude with a company annual report, a store-location list and your own case notes. Ask: “Map this company’s supply chain design choices, identify network trade-offs, list likely inventory risks, and generate five interviewer follow-up questions.” For broader operational transformation, connect this to automation and AI in operations consulting.
Interview Relevance
“A quick-commerce company is facing high delivery costs and frequent stockouts in two cities. How would you redesign its supply chain network and inventory policy?”
Say the trade-off aloud: “I would not optimise only for lowest logistics cost; I would optimise total cost-to-serve at a target service level.” That sentence signals maturity.
Common Mistake
The mistake: treating inventory as a generic cost to cut. It costs candidates because they ignore the service role of inventory and recommend reductions that create stockouts. Fix: reduce inventory selectively - by SKU class, demand volatility, lead time and service level - not uniformly.