Today was a day of reflection from this week of Kanban immersion training. While there are many information nuggets (such as using flexible PVC pipe to construct inventory racks) some of our most significant learnings are listed below:

There must be a stable process to have an effective Kanban system. If processes are highly variable/unpredictable a pure visual signal such as kanban is not effective. This is because the kanban is set on one cycle. The cycle works in a loop, so the loop must be in the same rotation over and over.

The role of the water strider (materials handler moving from station to station). This position allows the operator to focus on their core role/activites, while a resource dedicated to distributing supplies and equipment works around the operator to ensure they have what they need, when they need it.

Kanban at its core is visual. Everything about the kanban system is visually triggered. From the simple information on the card, to the limited flow pattern, everything about the process is easy to follow. The simplier the better.

A kanban should have a single purpose. The kanban triggers resources from point A to point B. It then returns to A for redeployment. It should be considered a round trip non-stop ticket.

Do not start work until a kanban is received and only produce what is requested. Kanban is a vehicle for standard work. If you start work prior to the signal, or work beyond what is required, you are contributing to the waste of overproduction (mira).

Placement of supermarket is key. Systems need supermarkets (inventory hold locations) to maintain flow between supplier and production points. It is critical to understand where the supermarkets are, what they hold, how resources flow through them, and how they are maintained.

Kanbans require continuous refinement. After implementation, assessments of performance become key. It is critical to ensure kanbans flow is controlled. Kanban replenishments should never be too “hot” (requiring many replenishments ) or “stale”, but should always be in synch with customer demand.

Establishing Buffer (safety) stock requires diligent study. Levels of buffer stock accommodate processes variability. However buffer stock should be defined with quantitative tools, reviewed with frequency, and established with purpose recognizing criticality and severity of the item.

Kanban does not create standardization, but it does rely upon standardized processes. Customer orders can be diverse (ie: different models and accessories on cars coming down an assembly line) but the items delivered to support the assembly process should come the same way every time.

Kanban is not a universal tool or solution. It does not fit all applications. It is a vehicle for flow but not always applicable. Sensei Niwa asked the question “Do you need a kanban card to tell you that you are hungry?” You want to develop systems that automatically pull the right supplies. Until you reach the point of innate pull within the system, the kanban can be the right answer.

The Supply Chain team has discussed plans for piloting a simple kanban system based on these learnings that can be started within the hospital and rolled out to the outpatient clinics. The goal is to remove the clinical staff from the supply replenishment process. This can be an iterative improvement on some of the early kanban systems that have been established.

This spring we will be working with our Tauber Institute (industrial engineering graduate students) interns to apply kanban approaches to some supplies in the OR core area and to the case cart process. Included in this work is possibly changing from the notion of case carts to case “kits” meaning that the OR is provided with exactly what is needed, for a case as it is needed. This would be a true kanban application.