2 Ways to Handle Varying Inventory Levels When Creating Value Stream Maps

Shaunak, a reader of LSS Academy, recently sent me the following question via email.

By the way, if you have ever have questions related to continuous improvement (lean and/or six sigma) feel free to email me.  I will do my very best to answer.

Hi Ron,

I have very quick question for you.

The level of inventory keeps changing all day, so when we draw the current state map, how to count the inventory?

I mean to ask: suppose SWIP decided is 20 pieces. At a point in time, inventory can be anything between 0-20.

Should we consider inventory level at the time when we are drawing the map or should we take the average inventory for the shift?

Smiles,
shaunak

Common Question

This is one of the most common questions I hear when teaching value stream mapping.

Specifically, if we draw the current state value stream map on Monday, and there are 42 pieces of inventory between stations A and B… but then I come back Wednesday and there are now 68 pieces… which do I document on my value stream map – 42 or 68?

Two Approaches to Handling This Problem

As with anything, there are several ways to approach this problem. Allow me to share two approaches I have personally used.

  • Document what you see on Monday (42) and aggressively move on towards your improved future state.
  • Study the situation over the span of a few days or weeks and note the inventory levels as both a measure of central tendency (mean or median) and measure of dispersion (standard deviation or range).

Mura Lives

Now, something else I’d like to mention is if you find yourself dealing with highly erratic inventory levels day in and day out you likely have other, more serious, problems besides not knowing what to draw on your value stream map.

Specifically, you’re likely dealing with severe case of mura (unevenness) due to a lack of heijunka (production leveling).

Don’t Agonize Over Details

So, instead of spending hours and hours, even days, worrying about what inventory levels to write on the VSM just document something and get focused on improving flow, preferably one piece at a time, while also working on finding a way to smooth and/or level production via heijunka.

You see, in the end, moving towards the ideal state is what lean is all about and this cannot happen if we never take that first, and often scary, step towards the improved future state.

How Would You Handle This Situation?

With this all said, I am curious on how you handle this situation.

How do you react to varying inventory levels on your value stream maps?  And for those in transactional worlds, even health care worlds, how do you deal with varying queue times?

11 Comments

  1. Jordan Davis

    January 7, 2010 - 11:04 am

    We tend to do your first recommendation but then follow up at least once a week and note the new inventory levels in the data box. We haven’t calculated the mean and range though so might give that a try. Could be interesting?

  2. Paul Cary

    January 7, 2010 - 12:46 pm

    The current state is a snapshot in time. The real power of VSM is envisioning and achieving a future state. In order to achieve a future state the mapping process magnifies the constraints or obstacles to flow (where inventory backs up or where there is no inventory). The current state is the current state, it’s the reconition of the constraints to continous flow that is important.

  3. S. Thomas

    January 7, 2010 - 12:56 pm

    We have recently taken a new approach to value stream mapping. We no longer draw a future state map. Instead we start with a current state map and then identify three things we want to improve. As we improve them we then update the map accordingly. We’ve found this to be very effective since before when we created future state maps our improved version rarely looked like we thought it should and p[eople would get frustrated.

  4. Bruce Baker

    January 7, 2010 - 4:03 pm

    I have generally followed your first recommendation. I have followed something closer to your second recomendation in cases where there was a long lead time on the upstream kanban loop – like a vendor with long lead time or a supermarket after a major ‘monument’ that serviced multiple value streams.
    Not related to the kanban part of the question, but I ask about making inventory levels below SWIP visual and escalating a situation where inventory is below SWIP to a problem.

  5. Ron Pereira

    January 7, 2010 - 4:08 pm

    Hey Bruce, I just checked out your about page on your blog and saw you are a Buckeye! I grew up in Greenville, OH and have loved the Buckeyes for as long as I can remember. What a game that was on the 1st, eh? Next year is the year brother. Believe that! Go Bucks!

  6. Harish

    January 8, 2010 - 7:00 am

    Time and time again, I see companies so muddled up in the art of process mapping called VSM.

    Let me ask you this – if you are a builder and you are assigned to replace an old worn down building, will you analyze the plan of the old building, walk up and down, look at all the flaws, spend weeks drawing out to the minutest detail what is wrong with the building

    OR

    Will you draw a new plan from scratch, understand the environment, and make sure the new plan is suitable and is what the customer wants, demolish the old building and build the new better building?

    Toyota used process mapping only during their TPS launch phase. The material and information flow diagram was later made so that they could train their suppliers on the concept – source Toyota Way. Why do you think, no Japanese books written by Ohno, Shingo or Monden talk about it?

    VSM is a great tool to teach new students on the concepts.

    If you are going to replace an old line, you are admitting that what you are doing right now is not working. Spend the time you would spend on Current, on the Future to make it happen. The concept of True North is to look forward and reach for what is possible.

    Ohno in Just in Time says – “In the plant we must first rearrange the equipment in the order that value is added to create the flow”. He did not speak about first doing a current state to see what all are we doing wrong. We are doing wrong. The cow has left the barn. No use closing the doors now.

    -Harish

  7. Ron Pereira

    January 8, 2010 - 8:39 am

    Interesting comment, Harish. Thank you for it.

    I actually agree with you that there are indeed times when people get lost in the current state and never actually make anything better.

    **BUT**

    I have also seen, as I am sure you have, the complete shock on people’s face when they create their first current state map of a process they thought was nearly perfect only to learn they had lots of improvement opportunities.

    So, while I think your analogy of the old building is valid in many cases… I also think there are many cases where people don’t realize that even though they don’t need a new building… there is definitely room for some renovation. And until they draw the map to see how poorly material and information actually flows they may not realize where to start.

    So, in my opinion, spending 4 to 5 hours (max) on the creation of a current state VSM can do nothing but help.

    At least I have never been part of a current state VSM creation process that has damaged the continuous improvement effort of the organization.

  8. Harish

    January 8, 2010 - 9:05 am

    Hi Ron,

    As I said, VSM is a great tool to train new lean practitioners. The only reason VSM is so used IMHO is
    1) Lean.org
    2) Consultants use it to train and start the lean implementation. This leads us to believe that we need to use it for every project.
    3) All lean practioners were trained in VSM first and they were trained to believe that VSM is a must for everything.

    My counter-argument always is – Let me tell you of a company that has been the BEST in lean implementation who did not use VSM to start their lean launch – Toyota.

    I have learned a lot from you and Jon Miller. I respect what you are saying. I consider Jon a great sensei, at least a virtual one.

    Boku no Lean no sensei o hontou ni sonkei shite imasu

    -Harish

  9. Bruce Baker

    January 8, 2010 - 9:34 am

    I wouldn’t dispute that Toyota didn’t use value stream maps when they started their implementation. I have heard some informed people suggest that Toyota’s rate of improvement between the late 40’s and the early 60’s did not exceed what many car companies did from the early 90’s to mid 00’s. I don’t know if that is or is nor true, I have heard that suggested as the truth. If that is true then it might suggest that Toyota is not genetically superior at learning and improving. They only adopted that philosophy earlier and that MIGHT explain why they are better at it than others. If that is the case then I would suggest that ‘because Toyota doesn’t do something, I shouldn’t do something.’ becomes a logical fallacy.
    I am a fan of Toyota and frequently hold them up as A model but never THE model. I drive a 1990 Toyota 110 miles each day. I am not disparaging Toyota only suggesting that Ohno, Shingo et al were men not deities.
    More to the point. In most cases I agree with Ron that 4 to 5 hours investment in the current state is a worthwile investment based on the assumption that there is usually some disagreement as to what the actual state is. I suppose that if everybody agreed and there was no ambiguity aroung where we are then we go immediately to work on where we want to get (maybe two levels – I like 10 yr and 18 months, I invite other viewpoints on how that, I am pondering how future should the future vsm be).
    If I can do what Harish suggests in his make a new building metaphor which is tear out my line (I can service customers otherwise) and rebuild it. Then I agree that the current state map becomes a purely academic exercise. The situation that I have usually found my self in is that I have value stream leadership that can’t agree oon where we are – literally -they don’t know what inventory is where, and I have to continue to service the customer so in my experience some amount of time (Ron’s 4-5 hours seems reasonable to me) is worthwile.

  10. Harish

    January 8, 2010 - 10:03 am

    Another analogy is hammer is a great tool. But you do not use a hammer for every occasion.

    I learn more from a process map and spaghetti chart than a VSM.

    The scenario that Bruce gave where the people do not where inventory is shows that they are not trained in lean. VSM is a great lean teaching tool.

    Let me rephrase – VSM is a useful tool. My only gripe is that it is used for every single project out there, and if it is not used then you are not doing lean. Lets take Kaizen Event for example. Kaizen event is a misnomer which generally means a kaikaku event. The difference is you do kaizen everyday and you do kaikaku frequently (monthly or bimonthly) as a team. Should we use a current state for these – most of the times I would say NO.

    A trivia – Shingo published the famous green book – A Study of the Toyota Production System from an Industrial Engineer’s viewpoint in Japanese, in 1981. Most TPS books I mentioned were written in late 70’s or 80’s. This was after TPS had already grown past its infancy stage to its adolescent stage. The Industrial Engineer that Shingo was did not include Material and Information Flow diagram in his book.

    Good discussion.

    Thanks,

    Harish

  11. Bruce Baker

    January 8, 2010 - 10:20 am

    I agree with your point that doing value stream maps very frequently like as part of an event or even with each project. It has been my experience that many, many events, projects etc. result from a single VSM episode. I don’t know that if there is not agreement where inventory is and how much means that people are not training. the maps that i speak of go from as far back in my value stream as I can go (maybe through one or two vendors) and go forward hopefully to end user. there are many people who understand lean but have not realized the opportunities in the ‘extended’ value stream. Many people like to map inside a plant which is probably worthwhile on some level but can potentially miss game changing opportunities. I don’t like to call my maps ‘extended’ because there is only one value stream for each product and it is a natural thing – all products begin in nature and ultimately end back in nature. The value stream is the whole path that materials and information take to do all that: extraction from nature, processing, transportation, storage, more transp, more processing, more storage, more transpo, more processing, distribution, installation (we can call this consumption), removal from installation, storage in landfill or recycling. Anything shorter than that is a ‘retracted’ value stream map. Mine arent’t extended. They are natural. I usually do stp at consumption when I am lucky enough to be able to see that part. I have yet to do one on reverse logistics.