Is the Kanban system to ensure availability of materials or to reduce inventory?

Pranay Nikam, from VCT Consulting India, asked the following question:

“I have designed and implemented the Kanban System at various type of industries. The challenge I face now is not that of explaining people how the system is designed or how it works. But rather clearing the misconception/misunderstandings key industry people have about Kanban.

My understanding of a Kanban System is ‘A Consumption based replenishment system’ with Multiple Re-Order Point (multiple Bins) as opposed to the traditional Two Bin System. In simpler words you keep enough stock to cover for the total lead time and add a buffer for demand variation and supply failures. And keep replenishing the stock as and when you consumes. The replenishment can be through fresh production or withdrawal from Warehouses or procurement from supplier.

Prime objective of the Kanban System is material availability to enable High Mix and low volume production; ultimately to support production levelling instead of running huge batches.

However, some Lean Consultants propagate Kanban as a inventory reduction tool and nothing more than a material scheduling software that can be configured in any ERP Systems.

I would be happy be receive your comments on the two different perspectives.”

The Kanban system has many variants, discussed in Chapters 10 to 13 of Lean Logistics. All these variants, however, have the following characteristics in common:

  1. They implicitly assume the demand for an item in the immediate future to match the recent past. It is a naive forecast, but hard to beat on intervals that are negligible with respect to what Charlie Fine calls the clockspeed of the business. And the fluctuations are smoothed by leveling/heijunka.
  2. They use some form of tokens to signal demand. Whether these tokens are cards or electronic messages, they can be detached from bins and parts and processed separately, in ways that are not possible, for example, in the two-bin system.
  3. There is a fixed number of tokens in circulation for each item, which is a key control mechanism for the supply of this item.
  4. The protocols for handling these tokens provide unambiguous directions on what should be done. No human judgement call is required to decide which item to move or produce. There are variations where that is not the case, like the French Kanban, which, for this reason, I don’t consider genuine.

The Kanban system  is not just a multiple-bin system, because bins are not used as pull signals. The Kanbans are pulled from bins when you start withdrawing parts from it, which you couldn’t do if the bin itself were used as a signal. If the signals are cards, you can organize them in post-office slots or on boards, which you also couldn’t do with bins. And, of course, you can do much more with electronic signals, which does not necessarily mean you should.

Your description of Kanban omits the goal of keeping inventory as low as you can without causing shortages, and experimenting with the numbers of Kanbans in circulation to test where the limit is, which makes it a tool to drive improvement.

Kanbans work for items consumed in quantities that have small fluctuations around a mean, which means medium-volume/medium mix rather than low-volume/high mix. You use other methods for different demand patterns, like reorder point for bulk supplies, consignment for standard nuts, bolts and washers, or just-in-sequence for option-specific large items… In low-volume/high-mix production you have many items that you cannot afford to keep around and only order from your supplier when you have an order from your customer; it isn’t the way the Kanban system works.

You can do many things with ERP systems but, historically, they have been more effective in managing purchase orders with suppliers than in directing shop floor operations. If you have an ERP system with accurate, detailed data about your shop floor, you can, in principle apply any algorithm you want to produce a schedule. Most ERP systems, however, do not even have structures in their databases to model the behavior of production equipment at a sufficient level of detail, and are not capable of producing actionable schedules. They print recommendations, and the final decision on the work that is actually done is a judgement call by the supervisor, or even sometimes the operator. Within its range of applicability, the Kanban system avoids this with simple rules, by focusing on what is actually observable and controllable at the local level.

So, I suppose the answer to your question is that the Kanban system’s immediate purpose in daily operations  is to assure the availability of materials while reducing inventory, with the longer-term purpose of driving improvement. Pursuing either of these goals at the expense of the other would be easier but not helpful to the business.

12 comments on “Is the Kanban system to ensure availability of materials or to reduce inventory?

  1. Dear Mr. Baudin,

    I appreciate the alacrity of your reply. Also, thanks to further bolster my understanding of The Kanban System.

    I would like to take this discussion in more depth.

    I agree that Kanban System does take into account the forecast for immediate future. The forecast for the next month or next quarter (depending on the nature of business) is averaged out for the calculation of the system.

    But the actual functioning of Kanban or rather operation of Kanban is triggered by real events i.e. consumption. It’s like our Principal & Advisor, Mr. Kumta puts it: “Plan as per forecast; Schedule as per Actual Demand”.

    The mention of Bins was just to distinguish a full-fledged Kanban system from the traditional two-bin system.

    Further, like you said, the bins may be used as a mechanism to act as a trigger but some times tokens or “Kanbans” may be employed too. The following is the picture of a Kanban we employ.

    The Kanban card (in the above link) contains a barcode, which can be used to electronically send triggers and also to register the receipt of materials. It contains all the information necessary to procure the given material.

    Kanbans can be employed for bulk supplies or for items of extremely low volume and frequency; what would change are the Kanban Quantity and the total number of Kanbans. The Runner-Repeater-Stranger analysis or XYZ analysis can help us determine the type of Kanban System to be employed. For instance: for an Item of very low consumption volume and extremely low frequency we have designed and implemented a One Card (Bin) Kanban system. So you order when you have consumed one entire bin. The bin remains at the location though, and it is the Kanban card, which moves to Kanban Board to initiate the trigger. Of course for unique items for which consumption may be very scattered say once in a quarter, the procurement has to be against order. Also in electronics industry where the technology rapidly changes, it can be dangerous to employ a Kanban System.
    Moral of the story, I believe that Kanban System can be employed for many situations provided there will be repeatability of demand for the given component.

    ERP is good for planning, but it fails to double up as a scheduling system. Accuracy on behalf of the BOM errors and entry errors; is one part and the second is the timing. Many ERP Systems are designed to update stock in a 12 or 24-hour cycle. This delays the generation of Triggers.
    Like you said “Most ERP systems, however, do not even have structures in their databases to model the behavior of production equipment at a sufficient level of detail, …… They print recommendations, and the final decision on the work that is actually done is a judgement call ……”

    But the thing that I can’t quite make peace with is the fact that Kanban is a solution to reduce inventory. Kanban System when scientifically designed and implemented at best optimizes inventory i.e. where it is low, it brings it up to a safe level and where inventory is high it bring it to a required level.

    If Inventory reduction as a direct impact is highlighted, the whole focus shifts to how the system can be operated using less and less Kanbans. E.g. knocking off the safety/buffer stock. In fact, at of my clients where Kanban was being introduced and we had covered almost 50% of the materials, people got into this habit of withdrawing few Kanbans of each item at the month-end to compensate for the increased inventory of non-Kanban Items.

    The focus should be instead on Continuous improvement and it should be directed towards reducing lead-time for replenishment. For e.g. Employing Milk Runs, eliminating the activity of packing & unpacking items by employing special withdraw bins etc. Increasing frequency of replenishment and cutting down the lead-time will result in smaller Kanban Quantities & Lesser Kanbans required; thus improving the Inventory Turns. Inventory reduction is therefore a byproduct and need not be pursued.

    • What I meant by forecasting was about the next pitch, not the next quarter. When you detach a Kanban from a bin and send it on its way, you implicitly assume that the quantity specified on the Kanban is what you will need when it arrives. Inside the plant, you assume, say, that the consumption rate 1 hour from now will be the same as it was in the past hour; with suppliers, the corresponding time scale may be one day. In an interval this short, the naive forecast is hard to beat. Otherwise, of course you are correct is saying that forecasts at the monthly or quarterly level play a role in planning.
      The only approach that schedules based on actual demand is just-in-sequence where, for example, you place the orders for the seats of a car at the time you start the car’s body on the first station of final assembly and the supplier delivers sequenced sets of seats to the station at which they are mounted in the car.
      The Runner-Repeater-Stranger analysis doesn’t just serve to determine what kind of Kanban to use, but how you should layout your production lines, supermarkets, and stores, and which items you should use Kanbans for, if any.
      The Kanban system is not a panacea. It is not applicable to everything, but mostly to Repeaters. For Runners, it is often overkill; for Strangers, its underlying assumptions do not apply. On rare occasions, at random intervals, you may receive orders for spare parts of obsolete models, and there usually is no point in maintaining stocks of materials and components for these spare parts.
      The rules of the Kanban system call for pulling cards when you withdraw the first part from a bin, not when it is empty, or half-empty, or two-thirds empty. Your sample Kanban contains no information about replenishment periods. A Kanban delivered to a supplier is not necessarily for immediate action. It normally specifies whether the parts are supposed to be delivered in the next milk run, two milk runs from now, of three milk runs from now, etc. so that the replenishment lead time closely matches the time covered by the parts in the bins when the Kanban is detached.
      Logically, XYZ analysis should be as useful as Runner-Repeater-Stranger, or ABC analysis, but the data needed for it are more difficult to retrieve and you can often make do without it.
      Milk runs are not really something you implement as an improvement after you implement the Kanban system but a prerequisite for its implementation.
      Rather than using extensive calculations, a common practice when implementing Kanbans is to start a loop with enough Kanbans to support the amount of WIP you initially have, and to whittle it down gradually by removing Kanbans from the system until you run into problems.
      Removing a few Kanbans from the system is something you do to expose problems and drive the kind of improvements you are describing, not to compensate for the increase in inventory of non-Kanban items.

  2. Dear Michel- I am not sure what you mean when you say “For Runners, it is often overkill”. Could you elaborate?
    Thanks Abhijit.

    • The boundaries between Runners, Repeaters, and Strangers are always somewhat arbitrary. If you sort your items by decreasing frequency of use, in the top ranks, you find items that go into every finished good you ship. In other words, if you are out of any of these items, you can ship nothing. These are the ones I call Runners. At the other end of the spectrum are items that are so rarely used that you can ship almost all of your products without them. If you can ship 95% of your volume without a particular item, I call it a Stranger. And I put all the items in between in the Repeater category.
      Among Runners, you have standard items, that suppliers can sell to many other companies, and items that are specific to your business. For standard items, commonly nuts, bolts and washers, it is simpler to use consignment than Kanbans. You let the supplier worry about your stocks, and pay based on the bill of materials of completed products.
      The consumption of the Runners that are specific to your business is based on your aggregate demand, which is more stable than the demand any individual product or product family. Because there are not many such items and a shortage in any one of them would idle the entire plant, you don’t mind keeping more inventory of them than you would for Repeaters.
      There are many different policies that you can use. You can have the suppliers deliver a fixed quantity at regular intervals that you periodically adjust. Or you can use Vendor-Managed Inventory (VMI). VMI differs from consignment in that you keep records in your inventory database for these items, to which the supplier is granted access, and the arrangements do not exclude paying the supplier directly for deliveries.

  3. Thanks for pursuing the debate. May I elaborate on a few points?

    “Sale OR Consumption at any point along the supply chain creates a pull on the upstream station. Thus consumption acts as a surrogate for Demand. Replenishment thus eliminates dependence on demand forecasts for deciding production schedules for the immediate future”

    Stocking Policy of spares completely depends on the Service Level one want’s provide to the customer. If we can’t afford downtime on a machine, stocking spares based on the past data and replenishing the stock either one unit or a set of spares when consumed is important.

    Along with the Usage Class-XYZ analysis, one has to take account of the Value Class-ABC of a component too. Runner (X) Item with a Low Value Class (C) is definitely a candidate of a simple Two Bin System where we can also factor in Bulk ordering. But a Runner Item with a High Value Class (A) is will not follow the same strategy; instead we aim for minimum inventory with a factor for variances and failures. Strangers by definition are out of the game whether they be of A, B or C Class. Value & Usage Analysis thus forms a basic screening for candidacy to a Kanban System.

    The release of a Kanban from a bin acts as a trigger for a buyer/shop floor supervisor to initiate replenishment. If the trigger is generated when the first unit of a bin is consumed, there is an inherent assumption that the whole bin will be consumed by the time the replenishment occurs. Whereas, consumption of an entire bin is the absolute truth, even if the demand for a month is fulfilled, under Kanban we only consider the next pitch, so irrespective of the calendar, the replenishment should occur.

    Kanban System can manifest into many forms. It is one of the most powerful scheduling systems ever designed. A simple Re-order Point system can be classified as a Two Card Kanban System and on the other hand we can link the assembly line to the foundry through the machine shop via interlinked Kanban Systems.

    The Kanban is a tool for continuous improvement. The metrics to be pursued are Lead Time and Replenishment Frequency. Unfortunately, the industry is driven by costs and in a bid to reduce them, they end up buying and shipping in as large quantities as possible meanwhile reducing the frequency. A Kanban System is a first step to show companies a new path towards cost reduction. Once implemented, the designs have to be revised at regular intervals depending on the clock-speed of a business.
    As much as we, the Lean practitioners would like all the pre-requisites for smooth running of a Kanban System to be present, we have to take the slow and painful path of educating our customers and then asking them to do away with their decade old beliefs. i.e. to bring about a Cultural change involving the entire workforce.

    • I am not sure who you are quoting in your first paragraph. However you cut it, the idea that you should order the exact amount you have consumed is an implicit forecast that this amount is what you will need.

      The relevant ABC analysis for production scheduling is not in terms of monetary value but frequency of use. A shortage in a 50-cent O-ring can prevent you from shipping a product as effectively as a shortage of a $2,000 engine. One may be more expensive to prevent than the other, and you may choose different methods on that basis, but you have to do both regardless.

      It is a basic rule of the Kanban system that it is detached when you pull the first part from the bin. I did not come up with that rule. You can check in all the Japanese literature on the subject. The loop is designed so that the bin is almost empty by the time replenishment occurs. If one milk run pitch is shorter than needed, you specify on the card how many pitches should elapse before it is attached to a full bin and delivered, and you place the card on the corresponding column of a board for action.

      The Reorder Point logic is different from the Kanban logic, although the material/signal Kanban combination used for stamping processes is close.

      The Kanban system is a good tool, but just one tool in a box of many. Don’t overestimate its importance. And it is rarely a first step. Usually, you have to redesign your production lines before you can make it work.

  4. Dear Michel,
    I am currently looking at investing into 2-bin Kanban for our hospital, all these vendors have different pricing models. What pricing model works best from a cost to our hospital, any suggestions. THe different pricing models are subscription based (where i pay an amount based on # of SKu’s handled), piece of the pie model (where the vendors get a piece of my savings by volume which directly ties to $), and third where its capital model.
    Can you suggest pricing model that would work best for me as customer and not just from a vendor’s perspective selling me 2 bin?

  5. The 2-bin system and the Kanban system are different. In the two-bin system, an empty bin serves as pull signal; in the Kanban system, this role is played by a card detached from a bin when you start drawing parts from that bin.

    For hospital operations, this raises several questions.

    1. Are you trying to use the 2-bin system or the Kanban system?
    2. What kind of materials are you looking to manage in this fashion?
    3. What kind of bins are you using?
    4. Do you have returnable containers circulating between your hospital and the suppliers?

    In the Kanban system, you usually have an general agreement, often called a “blanket purchase order” that specifies terms and conditions. Within that framework, each Kanban is an order for the amount marked on the Kanban, based on which you pay the supplier on the terms specified in the blanket PO.

    It is a model under which you pay by the bin you actually order, and the agreements sometimes include penalties if you repeatedly issue wrong forecasts to the supplier.

    If you use returnable containers, I believe it is best that you own them, because returnable containers owned by the supplier tend not to be returned promptly by the customer.

    Consignment or subscription systems are alternatives to Kanban, in which the supplier assumes full responsibility for the availability of the materials. In both cases, the supplier monitors the stocks of parts available where used at the customer site, and replenishes them as needed.

    In the consignment system, the customer pays on the basis of the quantities included in finished goods, based on the bill of materials. I am not sure how to translate this to hospital operations. For medication, the basis for payment could be the amount actually consumed by patients.

    In the subscription system, the supplier is paid a fixed amount periodically, and just makes sure that the parts are available to the customer. This is rarely used, and only for cheap commodities consumed at a steady pace, where the cost of monitoring closely exceeds the benefits of tighter control. I am not sure there are any medical supplies for which this level of laxity could be appropriate.

  6. Great article full of knowledge. I have some doubts regarding this process. I would be really grateful to you if you can clarify it.

    One of the company is using 1 Kanban card for replenishing 2 items. Now, they are implementing ERP and I am facing challenges because there can be only 1 kanban card for 1 item. Their idea is that when you submit that kanban card at store, store person will scan the barcode and replenish both the items at same place. Challenge here is that how to tie both the items together.

  7. For kits, just-in-sequence is more common than Kanban, because kits commonly contain option-specific items. Kitting is more difficult to justify is you use the same configuration all the time.

    In addition, kitting is best performed right next to and within sight of where the kits are consumed, which makes Kanbans unnecessary.

    On the other hand, if you do kit repeatedly used sets of parts and do it remotely, I don’t see anything wrong with using Kanbans to regulate their flow, with one Kanban attached to one kit or to a batch of kits.

    And you should not let an ERP system dictate whether you do it or not. If you need to do it and the ERP system does not support it, then don’t use the ERP system at this level of detail.

  8. That’s great piece of work sharing information and answering all the questions patiently.

    I have few questions regarding minimum quantity to buy and kanban System. I am right now implementing a 2 Bin system at my production facility. Actually it’s working fine and there are no shortages of parts as there used to be. But the main problem I am facing is minimum quantity to order and long lead times. How can I over come these problems and if I want to have less order quantity the price for each product is more and higher-level management is insisting to have MQO/EOQ to be ordered. They are thinking that they might fall short of goods moreover ordering in small batches costs me more. How can I show them the cost reduction in implementing 2 Bin with required quantity to order, they want to see results in numbers ($$).

  9. You need to examine the costs of small orders by comparing scenarios, such as daily milk runs to suppliers versus full truckloads from each supplier at a lower frequency.

    Assume you have a daily milk run to five suppliers and, at each, pick up small quantities that add up to a full truckload.

    Compared to running full truckloads to each supplier, the consequences are as follows:
    1. The workload is smoother for the suppliers, and they carry less inventory of their finished goods.
    2. The trucking cost is only marginally higher, if at all.
    3. You carry less inventory of incoming parts.
    4. Your inventory of parts is better balanced, avoiding the situation where you can’t produce even though you have racks full of all components but one and are out of that one.
    5. Part quality problems are detected sooner.

    Points 1 to 4 are reasonably easy to quantify a priori. You can measure the impact on quality after the fact.

Please share your thoughts: