A plant holds two weeks of finished goods as safety stock. The regional hub it supplies holds another two weeks. Each depot below the hub keeps a week and a half. Add them up and the network carries five and a half weeks of protection against one source of uncertainty, which is customer demand. In many such networks, three weeks would give the same fill rate.
Nobody planned that duplication. It grows because each location sets its buffer on its own numbers, with no view of the stock sitting one step upstream or downstream.
How Does Safety Stock Get Counted Twice?
Every buffer in a chain is sized against the variability it sees. A depot sees its customers’ orders. A hub sees the sum of depot replenishment orders, which arrive in lumps and swing more than the customer demand behind them. The plant sees hub orders, lumpier still. All three layers then react to the same underlying demand noise, amplified, and all three add cover for it.
Service targets compound the problem. Suppose the depot aims for 98% availability and the hub also aims for 98% on its shipments to depots. The hub is then protecting a level of service the depot has already bought with its own stock. Internal targets set this way rarely get questioned, since every single number looks sensible.
What Are the Signs of Duplicated Buffers?
Three patterns show up in the data.
- Total days of cover across all echelons is far higher than the longest lead time in the chain.
- Hub stock ages while depots still run short on the same items. That points to stock in the wrong layer, with too much held upstream and too little where customers order.
- Safety stock settings at different echelons were last reviewed in different years, by different teams.
None of these proves duplication on its own. Together they justify a proper network review.
How Should Each Echelon’s Buffer Be Set?
Start with the customer-facing layer, because only that layer has a real service promise. Decide what fill rate each depot must hit and over what replenishment lead time. Segment the items first. A fast mover sold at every depot behaves differently from a slow item stocked at two, and one rule should not set both. Then work upstream.
Each upstream stage gets a committed service time, which is the number of days it promises to take to fill an order from the stage below. A hub that commits to two days needs enough stock to keep that promise most of the time. Stretch that promise to five days and the hub can hold less while the depots carry more. The trade-off is real and worth pricing, since stock held further upstream is usually cheaper to hold and more flexible to redeploy. Downstream stock is closer to the sale. Choosing the committed times for every stage at once, with one service target at the edge of the network, removes the double count. Running that calculation across thousands of item and location pairs is the job of multi-echelon inventory management software, since spreadsheets struggle past a few dozen.
What Does the Fix Look Like on One Item?
Take an illustrative item with the five and a half weeks of cover described above. Set jointly, the buffers might split as one week at the plant, one and a half at the hub and about a week per depot. Depot fill rate stays where it was. Total stock falls by more than a third, and the freed cash shows up in the first quarter as old upstream stock sells through.
Why Does Supply Capacity Belong in the Same Conversation?
Buffers assume the plant can refill them on schedule. A line already running at its limit cannot rebuild a drained hub inside the assumed lead time, so the real replenishment time stretches and every buffer downstream comes up short. Constraint-based supply planning software checks the replenishment plan against what the constraint resources can actually produce before anyone commits to it.
Inventory settings and the supply plan belong in one review. New targets from multi-echelon inventory management software should reach the constraint-based supply planning software in the same cycle. The result is a network that holds less stock and can actually refill what it holds. Review one product family this month. Compare the sum of cover across echelons with the longest lead time, and the size of the duplicated buffer becomes plain. Where the sum runs to twice the longest lead time or more, duplicated buffers explain a large part of the excess.
Oritiq can walk your team through a network-wide buffer review in a demo built around your own locations.
