5 Questions to Ask Before You Attempt to Implement Kanban

My first manufacturing experiences were with two different multi-billion dollar manufacturing companies who happened to run massive ERP systems meaning we had massive MRP systems running the shop floor.  Sound familiar?

Anyhow, as a young lad out of college I just assumed this was how everyone operated. I also assumed everyone dealt with the issues of the “system” saying there were 150 widgets on the shop floor when, in fact, there were only 78.

As an aside, I once remember a conversation I had with a production control manager who assured me we had enough parts to complete a particular job. In fact he vehemently pointed to his monitor and said, “Look, we have over 300 parts!”

I was younger then, and less politically correct than I probably should have been, and went on to explain that I didn’t give two you know what’s what his computer said… I was just down on the floor and saw with my own eyes that we had less than 100 parts!

Go to Gemba

Well, after convincing him to get up out of his chair we went to the floor together and he saw that I was right.

I wish I could say we turned MRP off and implemented heijunka and world class kanban loops… but, alas, we didn’t.

It wasn’t until later in my career at another company that I saw how well kanban can work when implemented properly – the key words being implemented properly.

You see, I have also witnessed kanban failures and it’s not pretty.  Not pretty at all.

5 Things to Verify Before Implementing Kanban

With this said, here are five questions I would strongly encourage you and your organization ask before attempting to implement kanban.

  1. Do you have the necessary in house expertise? In other words, do you have someone on your team that has worked with kanban before? If not, I would encourage you to seek help from outside resources.  Hire some help or go visit some companies that are successful and do some intense benchmarking.
  2. Do you even need kanban? Here’s a dirty little secret… kanban is not the end all, be all to a lean system. In fact, if your company’s information and material flows smoothly from the moment you take the order until you ship and collect cash… rock on with your bad self! You probably don’t need kanban. Continue to focus on improving flow and reducing lead-time.
  3. Do you understand what heijunka is and why it’s closely connected to kanban? Attempting to implement kanban without first leveling and smoothing production is absolute suicide.  It’s like trying to run a marathon with a clamp on your nose and duct tape over most of your mouth.
  4. Is your production process stable? If you are producing defect after defect or your machines are constantly breaking down you must address this first.
  5. How are your changeovers? If you are to level and smooth production via heijunka (see above) quick and efficient machine changeovers are highly recommended.  So start here if needed.

Now, these are just 5 questions to consider. I am not pretending to assume this list covers everything… but it does include 5 questions you should definitely contemplate before attempting to implement kanban into your production system.

What do you think?

Do you agree with my list? Would you add anything?  Do you have any kanban success stories or, gasp, horror stories? If so, what went right? What went wrong?

 

7 Comments

  1. Irene Chan

    February 18, 2010 - 9:02 pm

    It has been learning experince for my company. We try to implemented kanban last year and it not work well at all. But after study some more we have started back slowly with one area and it has gone a lot better.

  2. Paul Cary

    February 19, 2010 - 6:13 am

    1) Who is the supplier?
    2) What is the suppliers lead time?
    3) How is the quality of the supplier?
    4) Is the suppliers on time delivery reliable?
    5) What are the constraints of the supplier i.e. min. lot sizes, transportation costs, shut downs, changeovers.?
    6) What container will the supplier ship the product in?
    7) How will the parts be presented when they are received?
    8) What is the daily usage based on recent history (6mo. to 12 mo.)( remove any exceptions from the calculation). Is there seasonality, if so how will you adjust for it?
    9) How many bins will be required?
    10) What quantity will be in each bin?
    11) What size and kind of bin will you use?
    12) Where will empty bins or kanban cards go?
    13) What information will be on kanban card?
    14) Will you train the workforce in the kanban system?
    15) Who will monitor the system?
    16) Will a master list be kept for all kanban parts?
    17) Will you review kanban items monthly
    18) Who will make the kanban cards?

  3. Mark R Hamel

    February 19, 2010 - 10:05 am

    Nice post! I think your second question should be the first one that is asked, preferably in the midst of value stream analysis and the development of the future state value stream map. Too often people don’t drive hard enough for continuous flow. They scare themselves out of pursuing it (ya, it’s hard)…or they sometimes mistakenly think that kanbans = lean.

    Supermarket pull should only be “plan B,” followed by sequential pull (FIFO lanes) – “plan C” or a hybrid of B and C. The reasons to pursue kanban (from a production standpoint) are things like shared upstream processes (among different value streams), unreliable upstream process, substantial changeovers, physical remoteness between upstream and downstream processes, etc.

    Lots to think about.

  4. Panu Kinnari

    February 23, 2010 - 12:26 am

    My first encounter with lean was during my thesis work when I was given task of designing and implementing kanban system in high mix, low volume-electronics final test and calibration line.

    In retrospect:
    1. Do you have the necessary in house expertise?
    Nope, one person who really had expertise left company a little earlier.

    2. Do you even need kanban?
    At that time I thought it was the way to go, now, not so certain.

    3. Do you understand what heijunka is and why it’s closely connected to kanban?
    Heijunka was part of my design, but I am pretty certain that importance of it was lost to line supervisor.

    4. Is your production process stable?
    Nope

    5. How are your changeovers?
    In minutes

    You probably can read from my responses that my kanban system wasn’t really a great success. Looking back I think the biggest reason for failure was lack of basic stability in the process. Immediatly after failed implementation and for a while after, I liked to place blame on line supervisor for not committing to My Plan.

  5. David

    March 24, 2010 - 8:27 am

    Relative to Kanban system for raw material I have a story.
    We implmented a very good and nice kanban system for raw material, even with short milkruns of half hour to deliver 1 hour stock to a central material spot near each line so a material hanlder was in charge to refill the line containers. Everything was nice but there was a big show stopper.
    The comitment from the people in line. for years they were “asking” for material according to their “tummy feeling”; they were asking for much more material than needed according to the half hour milkrun. It was hard but finally the people understood the system. Be aware that the people has a learning curve when you implement such a system.