Facility Location: Factors, Methods & Trade-Offs
Would you rather put one huge warehouse outside a cheap city, or ten smaller ones near impatient customers? That single choice can quietly decide delivery speed, inventory levels, freight cost, hiring difficulty and even whether the business model works.
- Facility location means choosing where to place plants, warehouses, stores or service nodes to serve demand at the best total cost-service-risk balance.
- The right location is not the cheapest rent. It is the best total landed cost after transport, inventory, labour, taxes, utilities, service levels and risk.
- Key factors fall into four buckets: demand proximity, supply access, operating economics and risk or regulation.
- Common methods: factor rating for qualitative comparison, load-distance for transport effort, centre of gravity for geographic placement and cost-volume analysis for fixed versus variable cost trade-offs.
- The core trade-off is simple: more nodes usually improve speed but raise fixed cost and inventory complexity.
- Interview answer structure: clarify objective, map demand and constraints, shortlist factors, choose method, compare scenarios, recommend with risks.
- The biggest mistake is optimizing one visible cost, usually rent or freight, while ignoring service level and network-wide cost.
Big Picture: Facility Location Is a Network Decision, Not a Real-Estate Decision
A facility is a node in a supply chain network. Its location affects inbound supply, outbound delivery, inventory pooling, labour access, capacity, resilience and customer experience. So the question is not “Which site is cheapest?” The question is “Which network design best supports the strategy?”
The Four Factors That Drive Facility Location
Think of every facility location problem as a scoring exercise across four factor families. The weights change by industry: a quick-commerce dark store overweights customer proximity; a steel plant overweights raw material and utilities; a bank branch overweights catchment demand and accessibility.
The Core Trade-Off: Cost Versus Service
Facility location is full of trade-offs. Centralising inventory in one or two nodes improves scale and inventory pooling, but may slow deliveries. Decentralising into many nodes improves customer proximity, but raises fixed cost, coordination effort and duplicated safety stock. This is why location decisions are usually evaluated through scenarios, not single-point answers.
This also connects directly to inventory policy. Once a firm adds nodes, it must decide what stock each node carries, how it replenishes, and what service level it promises. If you want to strengthen that link, revise setting inventory policy for a multi-product business after this topic.
A 2x2 Matrix: Four Facility Location Archetypes
Use this matrix when an interviewer gives you an open-ended location problem. First ask: does the business need to be close to demand? Second ask: is there a strong cost advantage in a specific region? The answer points to the archetype.
Methods to Evaluate Facility Location
In interviews, you do not need to build an advanced optimizer. You need to know which method fits which problem, what it measures, and what its blind spots are.
Worked Example: Centre of Gravity in 90 Seconds
Suppose a company must choose a warehouse location to serve three demand zones. Each zone has map coordinates and monthly shipment volume. The centre-of-gravity method estimates a location that balances weighted demand.
Formula: X* = Σ(load × X) / Σ(load), and Y* = Σ(load × Y) / Σ(load).
X* = (100×2 + 200×6 + 300×10) / 600 = 7.33. Y* = (100×8 + 200×4 + 300×6) / 600 = 5.67.
The approximate location is therefore (7.33, 5.67). In a real decision, this is only the first cut. You would then check roads, land availability, labour, taxes, lease cost, capacity and risk.
Metrics to Compare Location Options
There is no universal “good” number for facility location metrics because a pharma cold-chain network, an auto plant and a grocery dark-store network have different service promises. In interviews, define the target first, then compare options against it.
Definitions You Should Be Able to Say Clearly
- Facility location: choosing where to place operating nodes to meet demand at the best cost, service and risk balance.
- Centralised distribution: serving many markets from fewer facilities to gain scale and inventory pooling.
- Decentralised distribution: serving markets through more facilities located closer to customers.
- Total landed cost: the full cost of getting a product to the customer, including facility, freight, handling and inventory costs.
- Service coverage: the share of demand that can be served within the promised time or distance standard.
- Centre of gravity method: a weighted-average method used to estimate a location near the demand-weighted centre of shipments.
Case Study: Zepto and the Dark-Store Location Logic
Zepto shows how facility location becomes the business model when the promise is ultra-fast grocery delivery in dense Indian cities.
Zepto’s model depends on placing small fulfilment nodes, commonly called dark stores, close enough to dense residential catchments to make rapid delivery operationally possible. The strategic problem is very different from choosing one national warehouse. Zepto must decide micro-locations: which neighbourhoods justify a dark store, how much assortment each should hold, and how to replenish them without choking the network.
The primary driver is demand density within a tight delivery radius. A dark store only works when enough nearby orders can cover rent, labour, rider capacity and inventory holding cost. The supporting drivers are assortment discipline, replenishment reliability, rider availability, local real-estate fit and technology-led order batching or routing.
The lesson: in quick commerce, facility location is not a back-office supply-chain choice. It is the customer promise converted into geography. The winning network is not merely “more dark stores”; it is the right density of nodes, with disciplined assortment and replenishment economics.
How AI Changes Facility Location
AI is making facility location less static and more evidence-led. The decision is still strategic, but the data inputs are richer and the scenario testing is faster.
- Demand sensing becomes granular: ML models can estimate demand by pin code, hour, weather, festival period, basket type or customer cohort. This helps decide whether a neighbourhood can support a store, warehouse or micro-fulfilment node.
- Digital twins improve scenario testing: teams can simulate “what if we add one regional warehouse?” or “what if fuel cost rises?” before committing capital.
- Routing and replenishment data feed back into location: delivery time predictions, failed delivery clusters and stockout patterns can reveal that a current node is wrongly placed. For the inventory side of this loop, revise using AI for inventory optimisation and replenishment.
Use ChatGPT or Claude to create a location shortlist: paste a case prompt, define demand zones, ask for factors, weights, risks and a factor-rating table. Then challenge the output by asking, “Which assumptions would change the recommendation?”
Interview Relevance
“A D2C brand wants to expand from one warehouse to three. How would you decide the locations?”
Use the phrase “total network cost at the required service level”. It signals that you understand both operations and business strategy.
Common Mistake
Optimizing the cheapest site instead of the best network. Candidates often choose the lowest rent or lowest freight option and stop there. That costs them because facility location decisions affect inventory, service levels, resilience, labour, capacity and customer experience. Fix: always compare options on total landed cost, service coverage and risk together.