How to Properly Set Up e-Kanban in Manufacturing
2026-11-11 | 11 min Logistics and Manufacturing
e-Kanban must be configured so that the stock available at the moment a signal is triggered covers consumption until a usable replenishment arrives at the workstation. Key factors include the actual duration of the entire supply cycle, package size, and fluctuations in consumption. The electronic system must provide confirmation of signal receipt, allow for status tracking, and include a defined procedure for system failures. Inventory levels should then be verified at the specific workstation, accounting for potential delays and emergency deliveries.
A production line may have enough boxes at the start of a shift, yet two hours later logistics is arranging an urgent transfer. This is often caused by settings based on average daily consumption or travel time alone. However, between a box being emptied and new stock becoming available, there is also waiting for picking, the supply route to depart, and the material to be handed over.
Kanban is a signal that authorizes the production or movement of a specified quantity in a pull system. When supplying an assembly line, first determine whether the signal triggers replenishment from existing stock or orders production of another batch. These two cycles can take significantly different amounts of time.
The start and end of the cycle must be observable
Replenishment should begin with a clear event. This could be an identified box being emptied, a withdrawal from a designated location, or a defined quantity being reached. When several triggers are combined, a rule is needed to prevent two requests from being created for the same consumption.
The cycle ends when the delivered material is usable at the workstation. Confirmation that an item has been picked in the warehouse does not mark this point. If the delivery is waiting to be handed over, checked, or moved from a staging area, the line cannot use it yet.
The article on production supply and material downtime covers the basic link between consumption and the request. When setting up a specific route, this link should be supplemented by measuring every step all the way to the workstation. For one cycle, record:
- the time the signal is physically generated and the time the system receives it,
- the time spent waiting for preparation and picking,
- the time spent waiting for transport and the duration of the transfer,
- the time usable stock is confirmed as handed over.
This record helps distinguish slow picking from a signal that appeared in the system only after a delay. The corrective action can then target the right part of the process.
Consumption should be based on the production mix
The number of finished products per hour must be converted into consumption of the specific component. If the quantity of that component varies between product variants, the same line output can result in different material requirements. The order of variants within a short interval also matters.
The business therefore needs a time-based consumption profile for the planned mix. A daily average can hide a period in which several products with higher demand for the same part are made one after another. The supply route must handle that period or have an agreed procedure for servicing it.
Start-up, normal operation, and planned output changes should be assessed separately. Breaks must not be included in the average in a way that understates apparent consumption during active production. Similarly, distinguish a process-required allowance from unexplained discrepancies in the records.
Stock at the signal must cover the remaining replenishment time
For approximately steady consumption, a useful first estimate is: required usable stock at the signal = hourly consumption × replenishment time in hours + buffer. The buffer should cover a specific source of uncertainty. With uneven consumption, instead of using a simple product, assess expected consumption over the entire time window.
For example, if a workstation consumes 60 units per hour, 45 minutes, or 0.75 hours, elapse between the signal and usable replenishment at the workstation. The baseline coverage is 45 units. With an illustrative buffer of 15 units, 60 usable units must remain after the signal.
If the signal is triggered when one 20-unit box is emptied, the remaining 60 units amount to three more full boxes. In a simple loop where the just-emptied box is being replenished and the other three are available for consumption, that means four boxes in circulation. At the start, they may contain 80 units in total. This is a model based on a specific trigger, not a universal number of Kanban cards.
If the signal is triggered by removing a still-full box, or if a two-box system is used, the stock position changes at the time of the signal. The calculation must therefore follow the physical sequence. Do not count an empty box or stock that is still in transit as available coverage.
Variable replenishment time requires scenarios
The average cycle may look acceptable even though it regularly gets longer during part of a shift. A typical reason is the overlap of receiving, picking, and deliveries for several workstations. When designing the process, compare normal cycles with times during busy periods.
The planning time can be chosen based on the required reliability and a sufficiently large sample of measurements. A specific percentile or safety allowance cannot be set without data. If a longer planning time is already used, the buffer should not cover the same delay again without explanation.
As shown in the example above, extending the cycle to one hour with the same buffer means 75 units are needed after the signal. With whole boxes, this requires at least 80 units of available coverage. If there is no room at the workstation, the cycle could be shortened or the packaging changed. Simply adding more boxes may not fit the designed flow.
Packaging must match workstation and supply capacity
A large box reduces the number of requests but increases the time between signals and takes up more space. A small box creates more frequent tasks and may increase handling. The choice must take consumption, ergonomic access, and replenishment capacity into account at the same time.
Dimensions and weight must match the equipment in use and the approved workstation. Stock must not block access to material or displace the return of empty containers. For different package sizes, conversions between units, boxes, and handling units must be correct.
The design benefits from a broader view of material flow in the production warehouse. e-Kanban must fit into preparation, transport, and returns; otherwise, it merely shifts waiting to another part of the operation.
An electronic signal needs a closed lifecycle
A WMS can accept a request and create a logistics task if the specific solution supports this connection. Before configuration, define the request statuses and the events that change them. The operator must be able to tell whether someone is already preparing the replenishment.
Scanning the same empty box again must not automatically trigger a second delivery. Conversely, a new valid cycle for the same box must not remain blocked as a duplicate. Distinguishing the two requires the signal identity and its current cycle.
If the connection goes down, agree on a fallback notification and how it will be reconciled later. The worker needs to see whether the signal was received by the system or by a responsible person. A notification disappearing from the screen without confirmed receipt creates uncertainty, often resolved with another phone call and a duplicate trip.
An emergency delivery must preserve visibility of the original request
Urgent replenishment should be based on the time until stock runs out at the workstation and the actual feasibility of supplying it. If material is taken from a delivery already prepared for another workstation, assess the impact on that delivery. Every emergency trip should have an identifiable cause.
After an exceptional delivery, decide what should happen to the open standard task. It may be canceled, reduced, or completed based on current need. Without this change, both deliveries will arrive and stock at the line will exceed the planned limit.
When dividing ERP and WMS responsibilities, also agree who is authorized to change a request. An operational exception must preserve both the physical and recorded state.
Line replenishment setup worksheet
Complete these inputs for each material and workstation pair. The record should include the parameter owner and the date it should be reviewed after a change in the production mix.
| Area | Input or check | Decision |
| Consumption | Profile by variant and time window | Planned consumption during the cycle |
| Lead time | From signal to usable handover | Time and delay scenario |
| Buffer | Specific uncertainty and measurement evidence | Coverage without double counting |
| Packaging | Units, dimensions, boxes in circulation | Quantity and physical capacity |
| Signal | Trigger, receipt, duplicate, and outage | Closed lifecycle |
| Exception | Emergency delivery and original task | Responsibility for reconciliation |
The pilot should also cover peak delivery periods and a change in the production plan. Assess replenishment time, stockout frequency, and average usable stock at the workstation. Reducing stock only makes sense if supply reliability remains acceptable. Short review checklist:
- the mix, output, or component packaging has changed,
- waiting for preparation or transport has increased,
- lost signals or emergency trips are becoming more frequent.
Frequently Asked Questions
No. A consumption profile over the replenishment period is required. In the case of variant changes or an irregular production sequence, a daily average may underestimate periods of high demand.
It may temporarily create more time, but one must understand the cause of the delay and the capacity of the workplace. A lost signal or an unclosed request requires a process fix.
Following a relevant change in consumption, packaging, or the supply cycle, and based on the results of regular measurement. The review interval should be adjusted according to the stability of the specific flow.