In this lesson we’ll dive deeper into one of the “Massive Mistakes” topics: the design of formal delivery routes. But first, let’s review why this is important.
You already have delivery routes of some sort already. After all, materials need to get to their various Points of Use, one way or another. That’s not an option. What we’re discussing in detail here is the concept of the “Water Strider”, a delivery system that makes frequent deliveries with relatively light loads. Here are the main benefits that we discussed in the previous lesson.
First, frequent deliveries and light loads means less inventory at the Point of Use. Inventory reduction may be a goal, but saving physical space and improving operator productivity are equally if not more important. A significant “hidden waste” is the effort that an operator often needs to make selecting parts and moving to retrieve them. This can represent a big part of each cycle, and also introduce the opportunity for errors.
The philosophy of frequent deliveries also has the advantage of adding flexibility in the face of variable usage. If the usage of an item is faster than planned, for whatever reason, having frequent deliveries will help to overcome this.
One last thing before we proceed with the design details. Will a formal delivery route system increase your material handling costs? That depends on what you’re doing today, but it’s also the wrong question. The concern from a company perspective should be on the total cost, including operator efficiency, quality, and space utilization. The total cost to the company should go down.
Let’s now get into designing a delivery route, step by step. You’re going to need a few items to get started. First, get yourself a dimensionally correct layout of the delivery area. A CAD drawing of the area with a grid overlay will be very helpful to have.
Next, you need to create some standards: standard tasks and standard times. You need to define all of the repeating tasks that take place on a delivery route, and also estimate the time required to complete each task.
Here are some of the types of actions that take place, but feel free to add additional ones as needed: driving a tugger (time per yard), getting onto the vehicle, getting off of the vehicle, delivering containers, picking up empty containers, picking and refilling containers, and so on.
Estimate an average time for each task, and do actual time studies if necessary, to come up with reasonable estimates.
You now have a decision to make: whether to design your delivery routes as a decoupled system, or a coupled system. Here are the differences. In a coupled system the same person will both deliver the parts to the various Points of Use, and also be responsible for filling empty bins and picking parts. In a decoupled system, the functions of delivery parts and picking parts are done by two separate people. The delivery person will complete a cycle, and then immediately pick up the next group of containers or the next set of carts. When the delivery cycle is taking place, the second person is preparing the next set of bins and carts.
Which one is better? There’s no right answer, but larger facilities tend to operate in the decoupled mode, while smaller facilities will use the coupled method. Utilization of equipment with the coupled method will not be as good, since the delivery equipment will be sitting idle while the next load is being picked. On the other hand, it may be possible to respond more quickly with the coupled method, since the filled containers will usually be coming back on the very next cycle. This is not true of the decoupled method.
Let’s review what you have so far: you have a map of the area, you have standards defined for all of the potential tasks that will need to be done, and you’ve decided on a coupled/decoupled delivery strategy. What next?
You’re almost ready to start experimenting with different delivery routes, but you need to decide on a target cycle time for the delivery route. How many times a day do you want to delivery materials?
This is an important decision, for a number of reasons. More frequent deliveries will require a higher investment in people and equipment, and you may be used to thinking that material handling is “waste” that needs to be minimized. Frequent deliveries have a powerful benefit, however, that we’ve discussed previously. If you have variable usage, and who doesn’t, then relatively frequent deliveries provides a safety factor that allows the system to run normally, without expediting material and without excessive inventory quantities at the Point of Use.
Here’s a rule of thumb: try to set your delivery route time at about 25% of the average inventory usage time. For example, if you average a 1-day supply of material per Kanban bin, set your delivery route time to 4 times per day.
As a point of reference, Toyota Material Handling has set their delivery route time to around 35 minutes. This may seem too frequent, but they rarely run out of material on the line, and their inventory turns are excellent.
OK, I think you’re now ready to start designing your actual delivery routes. If you’re doing this by hand (or with a spreadsheet to help with the simple calculations), you will be trying out different logical routes through the facility in order to maximize your delivery efficiency within the time allowed. Use your spreadsheet to add up the distances and tasks proposed, and keep the total route time a bit less than your target delivery route time.
Feel free to experiment with alternatives, and if possible also get out on the factory floor and physically walk through the proposed routes.
This has been a high-level summary of the steps involved, and you can take it even further. Some companies will build a dynamic simulation model of the proposed routes, which will allow you to add variation to the standard times, and give you a better feel for how the route will perform. In most cases, however, you can simply design the route on paper, physically put it in place, and make adjustments as needed.
In the next chapter we’ll be talking about the creation and use of a “Plan For Every Part”, your database of everything you need to know about every purchased and manufactured item.
Overreliance on Kanban
Kanban is not dead. There has been, however, a sea-change in the way that the Lean community thinks about Kanban, and there is a growing openness to other material delivery options. Before I dive into that topic, however, a few words of introduction are offered, for those who need a quick Kanban refresher.
Kanban is a Japanese word for sign, billboard or more generally “signal”. In manufacturing the word is used to refer to a material delivery method that replenishes material based on a signal, usually a card, empty bin or an empty space. The beauty of the Kanban method is that material is delivered based on actual need, and the fact that previous material has already been consumed. This allows plants to control inventory tightly, achieve a high level of inventory turns, and virtually eliminate shortages. The Lean Mythology tells us that the method in manufacturing was inspired by American supermarkets, where small quantities of product are replenished based on actual sales. This did not escape the notice of visiting executives from Toyota in the 1950’s, and much of the original Kanban development is attributed to Toyota.
If you started your Lean journey in the 1990’s, Kanban was regarded as the material delivery method of choice in virtually every situation. For repetitive manufacturing companies (like Toyota) the method worked well. Other industries, like Aerospace or machine shops, struggled with the method, and instead used kitting and pick-list systems to deliver material to their consuming processes. A Lean consultant feels uncomfortable recommending kitting, since it seems to be a throw-back to an earlier, inefficient time. That way of thinking has started to change, however, even for companies like Toyota.
Keep in mind that all material handling is muda or waste. There is no such thing as value-adding material handling, since the handling by itself does not advance the product forward. You need to identify the most efficient methods for this necessary but non-value-adding work, both for the material handlers and also for the operators who are consuming the material. Here are five conditions under which a Kanban system may not be the most effective system for material delivery:
1. The material on the line is not in front of the operator. If the operator has to turn around, or walk some distance to retrieve needed parts, that is adding waste to his/her work.
2. There is a risk of error in selecting parts. It’s never a good thing to select the wrong part, but if items are indistinguishable with the naked eye, like different sized shims, then some countermeasure will need to be introduced to avoid selecting the wrong part. Multiple this opportunity to make a mistake across an entire line, and an error is just a matter of time.
3. The takt time is short. The shorter the takt time, the higher the percentage of the total time consumed by selecting parts, even if they are conveniently located.
4. High mix, Low Volume. If you are not consuming the same material consistently, your Kanban bins will remain full and unused, using both space and money.
5. Traceability Requirements. While it is possible to maintain lot control with a Kanban system, it is more complicated, since mixing loose parts from different lots in the same bin would not be permitted. Documentation would also be needed, to record the lot actually used.
What are the options other than Kanban? We have been conditioned as Lean practitioners to consider kitting as a dirty term, so we can’t use that word. How about the term “Kanban Sets” as an alternate phrase? In this method parts would be selected (not picked!) in accordance with the sequence of products being built. They would be delivered to the line via a cart or even an AGV, in small quantities on a Kanban Set Tray. This method can still be considered a pull system, since the signal to create a Kanban Set would be the arrival of an empty tray. It would not be necessary to have one set per product; the parts could be selected in small quantities to reduce the number of moves and the amount of handling needed. Line side material could be largely eliminated, freeing up significant space on the production line itself. The Kanban Set cell could be located logically in order to reduce the handling of incoming material. Pick to light systems, if desired, could be much more efficiently utilized, instead of using them on every workstation on the line.
The bottom line: under the right conditions a Kanban system is the material delivery method of choice. Under the wrong conditions, as listed above, a Kanban system can be a major contributor to inefficiency in labor, space utilization and quality. There may be a Kanban Set system in your future!