A team introduces kanban cards but continues to expedite orders outside the system. Containers disappear, quantities vary and suppliers replenish when convenient. The cards describe an intention rather than a functioning control.
Kanban authorizes work or replenishment within a defined system. Its usefulness depends on clear signals, reliable responses and controlled quantities.
Define what the signal permits
Specify the item, quantity, source, destination and action. A withdrawal signal and a production instruction can serve different roles.
Make the rule understandable: what triggers replenishment, who acts and how completion is recorded.
A digital task board may visualize work without implementing a replenishment limit. Be clear about which mechanism the organization is using and what behavior it should change.
Understand the replenishment loop
Measure demand during the time required to restore supply. Include relevant waiting, processing, transport and administrative delay.
Use consistent time units. If demand is expressed per hour, lead time must be in hours.
Size the containers deliberately
A basic calculation divides expected lead-time demand plus safety stock by container quantity.
| Hypothetical item | Value |
|---|---|
| Demand | 150 units per hour |
| Replenishment lead time | 0.5 hours |
| Expected demand during replenishment | 75 units |
| Safety stock | 7.5 units |
| Container quantity | 25 units |
| Calculated containers (82.5 / 25) | 3.3 |
| Containers in use | 4 |
Whole-container implementation requires four under those assumptions. Rounding down to three removes the stated protection. It should not be described as an automatic way to force improvement. Reduce the limit when evidence supports a shorter or more reliable replenishment loop and the consequences have been assessed.
Build the supporting conditions
Pull depends on inventory accuracy, appropriate container size, quality and the ability to respond to the signal. Frequent shortages can indicate a weak loop rather than insufficient employee discipline.
A hypothetical assembly line uses pull for stable components but plans long-lead engineered items through a different method. One control approach need not govern every item.
Segment by demand pattern and supply characteristics. Intermittent or highly uncertain demand may need another policy.
Reduce batches through improvement
Smaller lots can reduce waiting and inventory while increasing replenishment frequency. Setup time, transport and administrative effort determine whether that tradeoff works.
Leveling can reduce avoidable variation in the release pattern, but it cannot make actual demand uncertainty disappear.
Improve the conditions before repeatedly tightening limits. Protect customer delivery while learning how much buffer the system needs.
Keep the system observable
Track replenishment lead time, missing signals, shortages, excess and bypasses. Define how an exception is handled without silently abandoning the rules.
Review limits when demand, mix or supply performance changes. A successful initial calculation does not remain correct indefinitely.
Choose one stable item family and observe several complete replenishment loops. Compare actual conditions with the assumptions, then adjust the policy.
Kanban creates control when the signal, quantity and response form a repeatable mechanism that the team can understand and improve.