It’s Not the Formula that Matters


Kanban Calculation at Lean Sigma Supply Chain

The “Lean Sigma Supply Chain” blog has a post about the multiple variations of kanban formulas, the blogger writes:

“How can something so simple have so many variations. No wonder this tool has had so many failures.”

I doubt many kanban systems have failed because somebody had the wrong formula. I've used different formulas in different settings, based partly on the data that were available (or not available). It's not important if you calculate “I need 14.2 kanban containers” versus 14.32853. Kanban is not about having precise calculations.

If you have exact usage data for a part in the computer system, that's one source of kanban data (but the data might be wrong, so be careful about that). You can also use a survey of different users for a part, asking them “how much do you use on a typical day?” or “what's the most you might use in a given day?”

But having that initial kanban sizing calculation is just the first step. There are many ways that a kanban system can fail, most aren't related to the formula you used. Kanban will fail if:

  • There are no visual controls in place to indicate how the kanban system is performing (i.e., are cards being sent at the right time, are we stocking out, are we over-ordering?)
  • There is no standard process to make sure kanban cards or containers are being circulated properly, put in the right place, etc. (i.e., has everyone been trained?)
  • There is no management oversight to make sure standard work is being followed (i.e., does management formally audit the kanban system?)
  • There is no system for periodic review, for continuous improvement (or kaizen) of the kanban system (i.e., does the kanban system degrade due to lost cards or demand changes?)

It's not important that you have the most precise kanban calculation ever. It's important that you have a kanban system that works. Any thoughts? Have you struggled with any of the four points above? Do you use very quantitative data-driven kanban calculations or use estimates of part usage that you tweak over time?

Please check out my main blog page at

The RSS feed content you are reading is copyrighted by the author, Mark Graban.

, , , on the author's copyright.

What do you think? Please scroll down (or click) to post a comment. Or please share the post with your thoughts on LinkedIn – and follow me or connect with me there.

Did you like this post? Make sure you don't miss a post or podcast — Subscribe to get notified about posts via email daily or weekly.

Check out my latest book, The Mistakes That Make Us: Cultivating a Culture of Learning and Innovation:

Get New Posts Sent To You

Select list(s):
Previous articleIn Toyota-Land
Next articleTwo Sides to Every Story
Mark Graban
Mark Graban is an internationally-recognized consultant, author, and professional speaker, and podcaster with experience in healthcare, manufacturing, and startups. Mark's new book is The Mistakes That Make Us: Cultivating a Culture of Learning and Innovation. He is also the author of Measures of Success: React Less, Lead Better, Improve More, the Shingo Award-winning books Lean Hospitals and Healthcare Kaizen, and the anthology Practicing Lean. Mark is also a Senior Advisor to the technology company KaiNexus.


  1. All of those formulas seem more or less the same…figure out how many are used, how often they are replenished, how long it takes to get more, add in some safety stock and divide by the container size. It seems like they are trying to start a discussion over what color 5-S marking should be used to indicate an aisle way. Just pick something, pick an area to try it out, and go forward.

    This really ties in to a much bigger issue that many companies tend to gloss over when “implementing” lean. Kanban is not an inventory management system, it is a tool to highlight waste and identify problems. The reason the tool “has had so many failures” is because people expect it to be inventory control. In true lean thinking, it wouldn’t be considered a negative to have had many failures, it would be considered a success because it highlighted many problems that needed to be solved.

  2. Mark, got any data on how many kanban implementations fail to sustain? The trial and error approach is fine but not with hundreds or thousands of parts moving at high turns and customer service hanging in the balance. Getting the math right is important. Often, in my experience, folks won’t even get permission to start if they can’t explain the math. Again, in my experience, the greatest cause of kanban failure is the organization’s inability to recalculate kanban as conditions change.

  3. Hold on to your hats…

    More often than not I start with a KanBan of “1” (one) part between work stations/areas and see what happens.

    Most of the time it stays at 1, but more than 1 piece is sometimes justified by the processes.

    I could write a huge article on why I do this, but let’s just consider the concept of “1 piece flow” as a major reason behind it.

    Truth be told I also calculate a “Hybrid takt time” (in most cases,) based on what is possible to build, not what the customer will consume. Then I adjust staffing and resources accordingly in order to satisfy the customer.

    I know this may be upsetting to some and may raise a lot of questions, but Lean must itself be a Lean process and be very flexible in my humble opinion.

    All the Best,

  4. Larry, I don’t have any data. No you? All I have is personal experience, as well.

    You missed the point of my post. I said, as clearly as I could, that the getting the kanban sizing right is important. You don’t want to set it too low and shut down your assembly line (or operating room). But, my point about your post is that I don’t think people fail because they have the “wrong” formula. They might fail because they get the math wrong (they set a re-order point at 3 days when the supplier lead time is 2 weeks). Flawed thinking can’t be fixed by the “right” formula.

    My point was to say that having the math approximately correct is important, not to N decimal places.

    You can get the same kanban sizing by:

    1) Having precise usage data, with an average and standard deviation statistics

    2) Empirically determining “what’s the highest 3-day usage” of a given part?

    3) Estimating usage by asking 10 people.

    In some environments, such as a hospital lab, all three methods have given the exact same kanban answer.

    So, in my experience, the answer was very independent of the formula used. If you don’t have precise data, that doesn’t have to slow you down from starting with kanban. Pick a kanban size, err on the large side, and make sure the processes work and that you have a sustainable system. Then, tweak the inventory downward over time if you find out you need less safety stock, etc.

  5. Bill, thanks for your comment. You’re talking about in-process kanban, which is a whole different animal, eh?

    I think Larry and I were talking mostly about supply chain / supplier kanban pulls.


Please enter your comment!
Please enter your name here

This site uses Akismet to reduce spam. Learn how your comment data is processed.