Nine depots that each hold 100 units of safety stock carry 900 units between them. Put the same protection in one central warehouse and, if depot demand is independent and similar in size, roughly 300 units cover it. Pooling works because a high week at one depot tends to cancel a low week at another. The saving follows a square-root pattern, so nine locations need about three times the buffer of a single site. Real networks rarely deliver all of it.

Customers order from the depot nearest them, and stock sitting 600 km away cannot fill an order needed by tomorrow morning. Placement is a service question first and an arithmetic question second.

What Does Each Layer of Stock Protect Against?

Safety stock at any point in a network covers demand variation during the time it takes to refill that point. A central warehouse refills from plants or suppliers over several weeks, so its buffer absorbs supplier delays, production misses and swings in total network demand. Depot stock refills from the centre in a few days. It covers the short window between one inbound truck and the next.

Holding both buffers is normal. Waste starts when each one is sized as though the other did not exist. Setting every buffer together, through multi-echelon inventory optimization, removes that double count and shows what each layer needs once the layer above is counted.

When Should Stock Move Closer to the Customer?

Four conditions push safety stock out to the depots.

• The promised delivery time is shorter than the transfer lead time from the central warehouse.
• Demand at the depots moves together. When a seasonal peak lifts every depot in the same week, pooling saves far less, because the highs no longer cancel each other out.
• The item sells fast and steadily, so it turns quickly wherever it sits.
• Freight per unit on a depot transfer is high compared with the cost of holding that unit for a month.

Slow movers point the other way. Take a part that sells four times a year at each of eight depots. Held locally, it needs at least one unit at all eight. Held centrally, two or three units can cover the whole network, provided a courier reaches any customer inside the agreed delivery window.

How Much Does Correlated Demand Shrink the Saving?

The square-root rule assumes depots behave independently. Demand seldom does. A network-wide promotion, a price change or a run of hot weather lifts all depots at once, and the central buffer then faces the combined swing with little offsetting. Measured pooling benefits usually land somewhere between the full textbook figure and zero.

A simple count helps here. Pull two years of weekly sales by depot and count how often the depots rise and fall in the same week. Where they move together nine weeks in ten, the central saving is small, and a larger share of safety stock belongs at the depots. Where the moves look random, central stock earns its keep. Lead time variability matters too. A plant that misses its production date by a week now and then needs a buffer upstream, close to the source of the delay. Depot stock cannot fix a supply problem two echelons away.

Who Should Make the Placement Call?

Placement touches service levels, working capital and freight budgets in one decision. That makes it a topic for the monthly planning review, and a poor fit for one planner working alone in a spreadsheet. Good sales and operations planning software puts the inventory plan beside the demand and supply plans. Finance can then see the cash tied up at each echelon while sales sees the fill rate each option buys.

Bring two or three placement options to that review, each priced in stock value and expected fill rate, and agree the service target per customer segment before anyone debates units. A depot serving key accounts with next-day delivery deserves a different answer from one serving distributors who order weekly.

Where Should a Planning Team Begin?

Pick one high-value item family and map its lead times between each echelon. Measure how closely depot demand moves together, then price the options. Expect the answer to differ by family. Fast movers with correlated demand often end up mostly at depots, while slow, costly items end up mostly central. Put the placement call on the monthly planning agenda, where service promises meet cash targets, and revisit the split each quarter, because demand patterns and lead times drift.

Author Bio

Arvind Singh Rana is a co-founder of Oritiq, a supply chain planning platform for manufacturing organisations. Over more than 20 years in supply chain, Arvind has led over 100 transformation engagements.