I don’t usually share applications here as much since most if not all of them are being released into the library… but I have to make an exception for this one. Kanban!
I just love material systems. So much. Over the years I’ve just become very interested in how materials move through operations and it is fascinating to watch it run well (or to watch a stockout absolutely bring an operation to its knees!).
Anyway - the Tulip platform has improved to a point where Kanban can be very easily implemented with no hacks or workarounds. It’s about time we gave Kanban a fresh coat of paint! Honestly, I wish I still ran some manufacturing cells because I would use this almost immediately I bet. Maybe I’ll use it for items around the house?
Do you know advanced kanban system with accumulation/triangle trigger and FIFO launcher?
Can be the next Kanban app challenge!
@youri.regnaud believe it or not, that is exactly what the next step was going to be! Next was to add in accumulation (which I am think of, from a kanban calculation standpoint, as more of a multi-card reorder point). As for FIFO launcher I may need some clarification on that - do you mind elaborating? My intention was basically to have the accumulated kanbans trigger what amounts to a work order within the system. This single work order would be linked to these kanban cards and by fulfilling the work order the kanbans would be replenished. I was originally intending the work order to be placed in a backlog state upon creation, with a due date, and allow the scheduler in the area to determine how and when to schedule. For example, perhaps the planner has a setup wheel of some kind (maybe it is a paint booth?).
The next advancement after that was to make all of this dynamic kanban, because having to go in and change out your cards every time demand changes is a pain. Still working out how to do that exactly but it is definitely something I would want to put in place!
One of the problem with Kanban is the loss of replenishment priorities. The purpose of a laucher is to keep the FIFO order after accumulation is completed. the kanban can also be a source of production stress for fear of not having to produce anymore. The launcher makes it possible to visualise the production order load to be launched
In a digital Kanban system, accumulation can be “headless”, but launcher must be visible by production supervisor or/and operator
Perhaps one last comment. The number of Kanban cards is not dependent on consumption, but on the variability of consumption and the frequency of replenishment. Many approaches of pull-flow via the Kanban method fail because of empirical formulas. Contact me by email if you want more informations about formula what works
It sounds like we understand each other, but I love talking about material replenishment systems and it sounds like you enjoy them too! Maybe we can hop on a call to discuss, sometime.
A few notes in response:
Currently I am having the accumulation trigger a material request, which is a separate entity from the kanban itself. The due date is established as you might expect from the established lead time, and all accumulated kanbans will be expected to be fulfilled by the single date. This allows the multiple cards, in effect, to act as a reorder point.
In terms of FIFO, the material request will have a creation date that can be used for accurate FIFO, but it may instead be the case that a planner has optimized their operation using a setup wheel, so some requests may actually be processed before others.
From my experience, I’ve found that evaluating patterns of demands and patterns of supplier performance is the best way to size kanban loops. Highly variable lead times, for example, may result in an increase in safety stock, which would increase either the number of cards or quantity per container.
@youri.regnaud Here is a video of the kanban application with accumulation built in - per your feedback. How do you think the material request table and its created date field meets the FIFO launcher requirement that you’ve highlighted?
I think I’d like to add dynamic kanban, next, with some label printing, etc…
Hi, congratulations your apps rocks! What’s next? To perform analytics around to control your pull flow system and also better manage change of cards number, eKanban works better with disposable label/card. Each time our accumalation threshold is reached we create new empty cards (records) with unique number instead of change status of a existing record. With this fetaures, we can record all the key date by card (creation date, start in production, full, partially empty, empty, …) Another benefit of this design is the flexibility to increase or decrease the number of card in the loops. Each time you have a event, we can decide when/where add a card in the process
And if you master kanban you can try other methodology for High Mix Low Volume like QRM/POLCA or CONWIP as we live in a VUCA world!
Haha! Kanban is too fun. Anyway, this isn’t my app, really, it’s for anyone who wants it! For better or for worse I no longer have ownership of performance for any specific manufacturing operation. I try to make apps that have general utility among Tulip users and learn a bunch along the way.
As for the dynamic kanban, generating a new label every cycle and potentially changing to accommodate new parameters… that is definitely where this is going! Call me an environmentalist, but I really want to accomplish this without having to print excessive labels. I’ve built something like this for a machine shop in my past and I found that we wasted so much paper and the configurations didn’t change as often as I expected.
You would still print labels if you needed to add cards, but what if the QTY and daily usage weren’t printed? I know this is blasphemy when it comes to traditional kanban, but if we can reduce the paper waste then maybe we should. All that would be required to see the details of any given kanban ID would be to point your cell phone camera at the barcode. The item, supplier, location would all be displayed on the label but the QTY, Daily usage, lead time, number of card - being dynamic - would require you to look at the bin with your cell phone.
I bet it would work just as well this way, what do you think?
A simple way is to separate the notion of a reusable container with a fixed label (license plating) from the notion of a kanban card. This fixed label is linked to a virtual kanban card via the data model. With this approach, no label is thrown away while retaining the notion of a unique e-kanban card