Main content

Gray and Wallace

Chris Gray and Tom Wallace

Overheard in the halls of the Acme Widget Company:

Jane: I hear you've been having some problems meeting takt time in the large widget assembly cell. What's going on?

Bill: The problem is big volume increases from the customers. We'll probably need to redesign the cell and add people to get an engineered cycle time in the right range. Then we'll need to relevel the schedule.

Jane: What about using inventory out of the finished goods supermarket?

Bill: If it were just normal demand variability, we'd have enough in the supermarket to handle it. But it isn't—we're seeing a significant volume increase. We can keep the operational takt time at 65 seconds until the supermarket stock is gone, which will happen in less than a week. After that, we'd better have higher throughput in the cell or we're going to miss the customer call-offs.

Jane: How are things otherwise?

Bill: We're fine. We've dropped the EPE interval to less than a week in most of the shared resources and our heijunka process is working great. No, the major problem is in the assembly pacemaker cell.

Wait a minute. Who are these people? What are they saying? And what are they trying to do? They live in the brave new world of lean manufacturing. Just as lean manufacturing calls for processes quite different from traditional ones, it also has a different set of terminology. Let's see if we can get a handle on what Jane and Bill are discussing.

Takt time 
The basic rate of production—what Acme needs in order to meet customer demand—is the takt time, sometimes called the drumbeat for the process. Takt time communicates the frequency of demand and, consequently, the frequency at which a product must be produced by the finishing process.

The basic calculation of takt time is:

Takt time = operating time ÷ required quantity

For example, if Acme's customers require 240 items per day and the factory operates 480 minutes per day, takt time is two minutes, or 120 seconds. To meet customer demand, the plan must authorize sufficient resources, both in people and equipment, to produce one unit every 120 seconds. The cell or finishing line for the product needs a proven engineered cycle time of 120 seconds or less.

Using the language of demand and supply, we can say that takt time is the demandfor capacity and that engineered cycle time represents the supply of capacity. Supply must equal or exceed demand, or the total demand will not be met and the customers will be disappointed.

So takt time is the rate of production required to meet customer demand. This is the simplest approach, when conditions allow. Frequently, however, there are other issues that need to be considered in setting actual production rates. Inventory adjustments, products with seasonal sales curves, plant vacation shutdowns, intermittent large demand shifts, and other factors may require a wider view than pure takt time provides.

Operational takt time is the rate of production required to meet customer demand as well as the other factors cited above. Within the context of sales and operations planning (S&OP), pure takt time would be calculated from customer orders and forecasts; operational tact time would be derived from the production plan, which is the pure demand plus or minus necessary adjustments. In Acme Widget's case, the operational takt time should be used to determine the required output rates, taking into account the need to replenish and expand the finished goods supermarket—a term we'll discuss in a moment—because of the increase in demand.

The pacemaker process and heijunka 
A basic concept of lean manufacturing is to schedule at only one point in the overall value stream (the value stream comprises all the actions required to bring a product from order to delivery). Scheduling at this one point—the pacemaker—results in pulling work from upstream processes and flowing product to the customer through the subsequent processes. The planned volume and mix at the pacemaker process typically corresponds to what's known as the master schedule; the scheduled mix for the day's actual production, as defined by the heijunka box (see sidebar for definition), corresponds to the finishing schedule.

Pacemaker schedules are established in accordance with the takt time for all items that go through the pacemaker process. The entire system depends on a foundation called leveled production or heijunka. As Taiichi Ohno points out in his book Toyota Production System, "…production processes must be managed to flow as much as possible. This is really the basic condition." He goes on to say that another important condition is leveling production as much as possible.

Finished goods supermarkets 
Producing only to customer orders is, when practical, the best way to operate. However, producing to a small inventory may make more sense than flexing labor and plant processes to match day-to-day order variations. Products are often produced to a buffer called the finished goods supermarket rather than to customer orders.

For situations where the customer provides stable and reliable demand, it's often possible to operate with little or no finished goods supermarket. For a customer whose demand is highly variable, maintaining a larger supermarket will probably be required.

Supermarkets are essential for our friends at Acme Widget for two reasons. First, they provide a buffer of finished goods between highly variable customer demand and the pacemaker process that must run at a more stable and leveled rate. Second, they can decouple processes that run at different rates, for example a finishing process that flows at a constant rate and an earlier fabrication process with a long setup that runs large lot sizes at a much faster cycle time. These supermarkets do not contain finished goods but rather components that are used at the pacemaker (final assembly).

The size of an item's supermarket depends on the frequency of replenishment and the volume and variability of demand for the item.

Pull to the customer 
Another essential element of lean manufacturing is the principle of demand pull, which says "replace what is used."

Pull replenishment is quite simple. Inventory for items in a supermarket is divided into equal units called kanbans. As a kanban-worth of inventory is consumed, it is reported. When the number of empty kanbans hits a predetermined level, a signal is generated to schedule replenishment. In a company with low setups, small order quantities, and lots of manufacturing flexibility, one kanban of consumption might signal one kanban of replenishment. In other cases, several kanbans may need to accumulate before a replenishment is signaled.

To use the pull replenishment method rather than a more traditional work order method, there are some prerequisites:

  • Demand for the item must be relatively repetitive.
  • Lead times must be relatively short.
  • Components must be available so an item can be produced on demand when the visual signal is generated.

Of these prerequisites, the most difficult to achieve for most companies is relatively repetitive demand. Techniques used to create level demand include finished goods supermarkets and load-leveling mechanisms like the heijunka box, plus other level-scheduling methods for both volume and mix.

For items that are being replenished using pull, some of the traditional techniques often associated with material requirements planning (MRP) are shut off. Examples would include order releasing based on planned orders and traditional shop floor control based on push dispatching rules. Typically, however, MRP planning continues to run to project requirements for suppliers of purchased components and raw materials.

The issue is how to communicate execution signals. In the conventional production situation and for low-volume, intermittent usage items, order releases would typically be used. In the lean situation for high volume and relatively continual usage items, the best execution mechanism is almost always demand-pull (kanban). Here, a pull signal can indicate the need to replenish finished goods inventory, produce a product to order, and replenish component inventories.

EPE interval 
Long setups are an impediment to creating value streams that flow. They typically result in larger lot sizes, lumpier demands, longer lead times, potentially poorer quality, more inventory, and surges of work upstream. The large amount of time needed to change over a pacemaker may prevent a true mixed-model schedule and stand in the way of implementing effective pull systems for upstream components.

The every-part-every-interval (EPEI) shows the effect of current setups on the flow through a process, plus the frequency that each item produced in the process can actually be run without exceeding available capacity. The EPE interval calculates the capacity required for running all the items that go through a process and uses that to determine how much time is available for setup. The amount of time available for setup can be factored by the total setup time for producing every item at least once and from this can be calculated the maximum number of times you can set up every item and the order quantity for each one based on that frequency. For example, in the EPE interval calculation shown in Figure 1, Acme has calculated that each part going through the molding process can be produced twice each month—an EPE interval of 10 days for each part. This enables the company to set a lot size equal to ten days of demand for each item without incurring any kind of capacity problem and without running a larger lot size and more inventory than required. The company would typically do this same calculation for each different manufacturing process including its pacemakers.

The EPE interval answers the key question: How often can we run every part through the process? Once a month? Once a week? Once a day? Once an hour? The EPE interval validates that the schedules are actually doable from a capacity perspective. For fabrication processes that are not part of a pacemaker, the EPE interval is essential for setting lot sizes. Lot sizes based on the EPE interval ensure work scheduled for the process is doable and the most repetitive possible in the current environment.

The benefits of running each item using the smallest valid interval (and decreasing it to an even smaller one) include

  • Reduced lead time across the value stream
  • Increased flexibility and responsiveness
  • Reduced in-process inventory
  • Reduced space requirements
  • Improved quality
  • Increased opportunities to ship on demand
  • Fewer surges of work upstream.

For the pacemaker process, reducing the setup time and consequently cutting the EPE interval to less than the ship window (the interval between shipments to a customer, daily or each shift or every two hours) has another potential benefit. This enables the company to meet small orders as easily as large ones. For many companies, an EPE interval of less than a day means that any item can be run essentially on demand. For this reason, a goal of mixed-model scheduling is to get the EPE interval to less than a day, shift, or ship window.

The role of planning in lean production 
Key scheduling people in a lean environment—whether their job title is master scheduler or something different—typically are involved in the following activities:

  • Developing better strategies for dealing with highly variable demand. They focus on reducing variability by inventory supermarkets or through a pure finish-to-order or make-to-order strategy to dampen the impact of that variability on the plant.
  • Leveling the schedule for both the volume and mix. They create a plan for flow that supports the drumbeat of expected customer shipments and enables the smooth movement of work through the plant.
  • Monitoring customer order patterns and validating the daily execution mechanisms. They produce to customer orders whenever possible at the exact day's mix using heijunka boxes or other load-leveling techniques.
  • Driving improvement activities so all processes can produce smaller quantities at shorter intervals. They create a true mixed-model schedule as well as more repetitive demand for components being pulled from upstream processes.

Such activities are in addition to these employees' traditional roles of balancing supply and demand, ensuring customer responsiveness isn't sacrificed in the name of factory stability (or vice versa), deciding where and how to meet the customer, and deciding how to manage changes to the plan.

These, then, are the basic tools needed for effective planning and scheduling in lean manufacturing: takt time, operational takt time, scheduling at the pacemaker, supermarkets, heijunka/load leveling, kanban/demand pull, and the EPE interval. They are also the most important areas to understand for those planning to integrate their master scheduling processes effectively into a lean manufacturing environment.

Will you and your company be moving to lean manufacturing? Our answer is yes—it is simply too good and too powerful to ignore. Some companies embrace lean 100 percent and reap great rewards. Others use lean on their high-volume products and remain in a conventional mode on the others. Some companies use lean principles for waste elimination but have difficulty with the scheduling elements of lean because of highly volatile, nonrepetitive demand and perhaps a broad and fragmented product line.

Almost all companies should be using lean. Too few are using it today. Lean will probably touch your life, if it hasn't already, so start building your knowledge base.


"Master It" originally appeared in "APICS - The Performance Advantage", the monthly magazine of APICS (www.apics.org).  

 

Chris Gray is the president of Gray Research and can be reached at cgr...@grayresearch.com . Over the last twenty years, he's helped more companies sort out manufacturing and distribution software issues than any other individual in the field and has authored three books on this topic. His Web site is www.grayresearch.com

Tom Wallace is a Distinguished Fellow at Ohio State's Center for Excellence in Manufacturing Management.