Kanban Replenishment: When the Simplest System Is the Best One

Kanban is, at heart, a signal: when a stored quantity of goods is consumed, a card or an empty container shows that it must be refilled. Instead of forecasting consumption and calculating an order, consumption itself 'pulls' replenishment. The logic is inventory control based on actual usage, with a tool immediately understandable to a warehouse or line operator.

Simple math sits behind the number of kanbans. The number of cards (or containers) equals consumption over the replenishment lead time, plus a safety margin, divided by the size of each container. If daily usage is 100 units, replenishment lead time is two days and each container holds 50 units, you need 4 containers to cover usage plus safety containers. That number also directly caps inventory, because you cannot produce or buy beyond the number of containers.

Kanban works well under these conditions: fairly stable, repeating consumption, low item variety, short and predictable replenishment lead time, and a short physical distance between consumer and supplier (internal or external). For standard items, fasteners, packaging and line consumables it is one of the simplest and cheapest systems.

But it does not suit items with highly variable, seasonal or project-driven demand, or long, volatile lead times, because a fixed kanban count cannot cover the variability except by expensively enlarging the safety margin. In these situations the number of kanbans should be reviewed periodically, or other policies such as a reorder point with calculated safety stock should be used.

Kanban needs discipline. When an empty container is not placed in its designated spot, or a card is lost, the signal is lost and the system goes dark. A designated place for containers, a rule of 'no production or purchase without a card', and regular audits are conditions for survival. The system should also signal if containers are always empty or always full: the first means too few, the second too many.

A common mistake is setting the kanban count once and never reviewing it. Consumption changes and the count must follow, say each season. Another mistake is using kanban for unsuitable items out of a liking for simplicity. Use kanban for stable class A and B items with high usage and for low-value, high-usage class C items, and manage the rest with other policies.

Electronic kanban (card scanning or sensors) delivers the same logic with fewer errors and tracking, but it need not be there from the start. Begin with physical containers and cards, and when the pattern is stable upgrade to the electronic version.

A practical starting point: select a group of 10 to 20 low-value, high-usage, stable items with short replenishment lead time. Calculate the kanban count for them, mark the container locations and run it for a month. Compare the number of stockouts and average inventory with the previous month.