Work in process (WIP) limits is a powerful, lightweight tool to not only improve your process flow but also to find further improvements in your process. I consider it widly underused but hugely impactful.
Often when WIP limits are introduced we miss the point of them being the driver for further process improvement, but rather focus on what our WIP limit should be, or how we are going visualize it on our board. So I often share a story on how that can work.
I realize that I’m turing into an old man… I have, for many years now, being telling and retelling the same story so many times that people around me don’t stop me anymore.
At the same time I sometimes forget some of those stories. So I thought I’d better write them down before I lose it altogher.
The door mounting guy
Imagine that you are working at a lean car factory. Your job is to mount the doors. The cars arrive in a steady stream down the line and you mount doors from a pallet that get to your station. The pallet has 10 doors on it and you pick a door at time to mount onto the card using an awesome looking robotic arm.
As you picked the first 5 doors you notice that the sixth door has a little card attached to it. That card, my friends, is a kanban card. It says:
Go over to door-builder station 14 and ask them to build you 10 more doors
You walk over there. Give them a highfive and then ask them to build 10 new doors. After some banter about the fading sun-burn in your face you walk back to your station and continue to mount doors.
Just as you mounted the tenth door, your turn around the the door-builder from station 14 comes by with a new pallet. Timed to perfection. It’s almost scary.
As this happens a lean sensei walks past you two and asks:
How is it going?
Everything is working fine here. We have door work in process limit of 10 and that is beautifully timed so that I get the doors just in time and don’t have to wait.
No overproduction or other waste here, sir lean sensei, sir!
He smiles and nods and then say:
Awesome. Let’s try to use 8 doors in our batch size.
What?! No! What did he just say? You oppose him:
No, wait… that’s not really good. Because then I will go over there after 4 doors and then they will have to work faster. I might have to wait until they are done. Or I get the doors early
Station 14 guy agrees:
Totally! I will have to talk with my suppliers to get smaller things faster from them and …
You both look up to the ever smiling sensei:
That is right. Fix those problems and we get an even faster, even smoother flow. We’re getting closer to our vision of one piece continuous flow. I know you will make it. Let me know how I can help.
Comment on car mounting guy
They use the WIP limit to provoke new problems to appear. Ah, that’s a bad word, problems. To expose new unrealized improvement opportunities for further improvements in their flow.
Notice also how the sensei is using the vision of the company (one piece, continuous flow) as a guiding north star for the improvements the local stations is making.
Are you continuosly pushing WIP limits down? Or are you being comfortable in the WIP you have now? Maybe a bit too comfortable? No problems is a big problem for a lean company.
A WIP limit is only best so far. We are always looking for better ways. By continuously decreasing it we will expose new problems. Hey, if we get to many problems, we might raise the WIP limit.
That’s the topic of the next story.
More WIP here, please!
This story I do know the origin from. It’s from the awesome Toyota Kata book by Mike Rother. Story is maybe exaggerating it a bit… it’s just a sentence.
A company invited a Toyota sensei to their factory to get help. As (s)he walked around the factory comments were made about different improvement opportunities or disturbances in flow, by the sensei.
At one point (s)he exclaimed and pointed to an empty machine:
More work in process is needed here!
That made everyone in the company burst out laughing. Obviously a joke, right? No - it was not. More WIP was needed. The process could not sustain such a low level of work in process.
Comments on More WIP please
Most likely, I was not there, the station in question was starved for new work. This happens on a visualized board when one or more columns in a row have no stickies in them. Meaning that if the downstream processes need more work they have no more (prioritized) work to pull.
This is not good either and can be viewed as a inverted queue.
The trick with WIP is that you want to balance it to a point where the process is flowing smoothly. Relating back to the previous story, this is not an end state but just a stepping stone to greater process improvements.
The other day I saw a board for about 4-5 people with 2 stickies on it, the column before at the end of the process. There were about 4 columns before those stickies. Meaning that there will be quite some time before new work reach this point.
Is this good or bad? That cannot be deduced from this (and my) knowledge, but I would rather agrue that this is a point where more WIP is needed to flow more work faster through the process.
Remember the wise words of Dr Reinertsen
Value trumps flow. Flow trumps waste elimination
Lower WIP is generally better, but other than that I leave you with my rule of thumb and simple heuristics for finding a good WIP limit.
Rule of thumb
Stop starting, start finishing
Yes from Kanban In Action . Make it a policy in your team. We prefer to finish things rather then start new things.
If WIP is too high - work will be idle. There will be work that people will not have the time to attend too. “I didn’t get around to that one” - comments in the morning meeting, for example
If WIP is too low - workers will be idle. Then you will hear comments like “I have nothing to do after this”. This is not bad in itself, since we now have slack that we can use for improvements, collaboration or lower priority work. If it happens all the time you should consider increase the WIP.