Kanban implementations are often disappointing and short lived. Companies are promised better service, lower inventories and level demand. But more often the system is abandoned after a short period as it has turned the shop floor into chaos… What went wrong?
We should understand a kanban is nothing more or less than a visual sign, typically a card. It is often used in just-in-time production, one of the pillars of the famed Toyota production system.
The just-in-time approach aims at making only what needed, when needed and in the quantity needed. It’s higher purpose is to serve our customer in time, without excess inventory due to overproduction (also wasting capacity) and with leveled capacity utilization. JIT principles are: to let the customer pull, to move towards one piece flow and to level demand.
Within the JIT system, kanban have three functions:
- Kanban as a communication tool to materialize the just-in-time principle of pull.
- Kanban as a way to create and develop autonomy regarding JIT at the gemba.
- Kanban as a visual control to see waste and to trigger Kaizen in teams.
However, kanban in itself does not necessarily lower inventories (it brings them under control in statistical terms) nor does it necessarily improve service. I’m not saying it wouldn’t, just not necessarily so. Kanban, as so many other elements in Lean, in the end is only a part of a more comprehensive system.
So how do we then actually ensure the kanban system delivers all on its promise regarding service, inventory and utilization, as well as providing autonomy and triggering improvement after it has been put in place?
Kanban: Surfacing our Problems
To understand what is required beyond kanban as a communication tool to materialize JIT, we need to look at kanban as a visual control.
Three potential issues related to the goals of JIT can be revealed on the gemba, by the team, through kanban:
- Parts are late, leading to poor service. This can easily be detected through missed picking at the supermarket and lots in a red launcher of an internal client. And when not yet late we can detect service risks through the absence of inventory or being below a minimum level, and lots in a red launcher of the internal supplier.
- Parts are early, leading to (possibly excess) inventory. This can be detected by being over maximum inventory, by parts without kanban, seeing products outside zone or parts that are available before their target date.
- Non-level utilization, which can be detected by seeing both empty and full launchers at parallel or sequential stations, seeing unbalanced inventories in supermarkets, having an unbalanced number of kanban in lot making boards, observing intervals between lots of the same article in a launcher that are not equal and not consistent with the EPEI of the part, observing that cards are not evenly spread around the kanban loop, and seeing either late or early picking for the staging areas.
Kanban: the Engine of Kaizen
The question now is, what do organizations do when confronted with the problems that kanban has highlighted? There is a serious risk that organizations abandon kanban within weeks or months after it was deployed when these problems are not addressed. Problems that were not visible and felt on the shop floor now suddenly are. It can lead to a “blood red” gemba I can assure you. So what to do?
1st Level Abnormality Management: Standard Reaction
As mentioned, kanban act as a visual control providing visual cues of problems in our material flow. Kanban are the physical link between the JIT and jidoka pillar in the TPS. They allow detection just like andon and poka yoke.
When, through kanban, we detect a problem, we need to share and contain the problem. A first level response is required in the form a standard reaction that allows the team to contain the problem, secure the customer and to prevent immediate escalation. In 8D terms it can be seen as the immediate corrective action or ICA.
Examples can be to use safety stock, to increase the number of kanban, to re-affect material of another order, to provide a visual cue to the supplier using missed picking kanban and to put the delayed lot in a red launcher and to prioritize it when the problem has been corrected.
We should be conscious though, that standard reactions as part of abnormality management do not address the root cause of problems. They only standardize problem correction. Therefore, we need more when we want to structurally improve our process. This where routines as part of Leader Standard Work enter the stage.
2nd Level Abnormality Management: Routines and Kaizen
The visual cues provided by the kanban system (like missed picking kanban and red launchers) allow Team Leaders and first level managers like Supervisors to detect areas and teams that are in need of support and improvement in order to serve their clients.
After problems have been prioritized and allocated, teams then move on to analyzing the problem and implementing counter measures that will permanently eliminate the underlying root causes using a structured problem-solving process like 8D. This might involve, for instance, smaller lots after reducing changeover times, reducing the number of kanban after a machine has been made more reliable or developing a more level supply pattern together with the customer.
When teams do not have the competencies required to eliminate the problem or when the cause is outside their area of responsibility, the problem can be escalated using the routines.
But what often is missing in kanban deployments are precisely these managerial routines. This endangers the link between the JIT and jidoka pillar in the TPS, making the whole system unstable.
When we don’t detect and address abnormalities in the kanban system, delivery will not improve, inventories will not be stabilized and reduced and capacity will continue to be wasted.
Routines enable management to learn about obstacles that are in the way of performance. Without them management won’t learn. They will continue to be puzzled about why performance isn’t improving and what they really need to do about it.
But not only that, very quickly the kanban system will be abandoned by the teams as they see that problems revealed by the system are not being addressed. At the gemba they will only see the blood red of all problems and they will blame it on the kanban system. They will not see the higher purpose of the system, and they will not feel empowered.
Routines are where purpose is given to the kanban system. Otherwise the system will remain a “game of cards” for many on the shop floor. Abnormality management and routines are where we use the kanban system as a visual control to ensure we meet our objectives, we jointly identify waste and opportunities for improvement, we show we respect our associates by enabling autonomy, kaizen and allowing them to deliver a meaningful contribution to the purpose of the company and where we turn around the hierarchy enabling the shop floor to escalate issues they now can see to their leaders.
But First, Respect the System
When the kanban system surfaces a problem, we need to be sure we are looking at a real problem of course. But when kanban are not correctly circulated, move late or early or even get lost, when lot sizes and the FIFO sequence are not respected, when visualizations are inconsistent with reality, or, in general, standards and standard reactions are not respected, the kanban as a visual control stops to function properly. We will lose the link between the JIT and the jidoka pillar in our production system.
As a result of these “adherence” issues we will see false negatives and positives. Due to our ineffectiveness to respect the system we may cover up real problems (as we don’t see them anymore) and at the same time may indicate problems when in fact there aren’t.
This is why I advocate to also make use of a management routine, also part of the system of Leader Standard Work, to review the degree to which the system is (and can be) respected by the teams. Management’s first focus should be on this aspect if it wants to use kanban to detect problems and to use it to drive kaizen.
Teams have a responsibility for respecting the system. If teams want respect in the form of autonomy, empowerment and opportunity to do kaizen, then inversely teams should show respect for the standards that make up the system. In that sense it is a two-way responsibility. Of course this respect is greatly improved if the system was set up with heavy involvement of the teams, proper training and ongoing coaching.
These routines also provide ample opportunities for leaders to actually coach their teams on the “why” of the system and to clarify how the visible nuts and bolts provide very concrete cues for a purposeful contribution by all involved.
Leader Standard Work for Respect and Kaizen
Therefore, whenever deploying just-in-time, pull flow and kanban, we need to both (1) secure the use of kanban as a visual control and (2) to rigorously detect and address problems identified through the application of kanban.
Management needs to do both. Not addressing the respect for the system will make the system foggy; not addressing the problems will make that the system doesn’t deliver on its promises.
It’s your time to play now: clear the fog and face your problems… Head on!