Replenishment: choosing the model that refills stock

Continuous review: order when the position falls to a trigger

Stock is monitored against a trigger level, and when the available position falls to it an order of a set size is placed. Because the trigger can be reached at any time, the buffer only has to cover uncertainty during the supplier's lead time. This model is efficient in stock terms and suits items where the position is reliably known, but it generates orders on no fixed calendar, which complicates consolidation with other lines from the same supplier.

Periodic review: check on a cycle, order up to a target

Stock is examined at fixed intervals and topped up to a target level, so order quantities vary while timing is regular. The buffer must now cover uncertainty across the lead time plus the review interval, because nothing will be ordered between checks, which makes this model more stock-hungry. In exchange it produces predictable ordering days, which is what allows several items to be combined onto one consignment and shipping economics to be captured.

Consumption-driven refill

Pull systems replace what was used rather than calculating what will be needed: a signal is generated as stock is consumed and authorises a fixed replacement quantity. Two-bin arrangements and card-based signals are the traditional forms. They work extremely well where consumption is steady and lead times are short and reliable, and poorly where demand is lumpy or supply is slow, because the signal only appears after the demand has already occurred.

Matching model to item

Fast, steady, high-value items reward continuous review and tight parameters. Slow or erratic items are usually better served by periodic review with a longer cycle, or by ordering only against confirmed demand. Low-value items with cheap storage justify simple pull rules and generous buffers, because the administrative cost of precision exceeds the stock it would save. The segmentation that drives this should be revisited as items move through their life cycle.

Where replenishment quietly goes wrong

The common failures are structural rather than mathematical: parameters set for volumes that no longer apply, an inventory record that disagrees with the shelf so the trigger never fires correctly, order quantities inflated by supplier minimums until they dominate stock, and manual overrides that have become permanent. Auditing a sample of items against their actual demand and lead time usually finds more value than adjusting the model itself.

Frequently asked questions

Why does periodic review need more stock than continuous review?

Because the exposure window is longer. Under continuous review the buffer covers only the lead time; under periodic review it must also cover the interval until the next check, since demand arriving just after a review cannot be responded to until the following one.

Can different models run in the same warehouse?

Yes, and they usually should. The model is an item-level property, not a site-level one, so a facility can run continuous review on its fast lines, periodic review on the tail and consumption signals on consumables without any conflict.

What breaks a replenishment system fastest?

Inaccurate stock records. Every model reads the current position to decide, so when the record disagrees with reality the rule fires at the wrong moment in both directions, producing simultaneous shortages and excesses that look like a planning failure.