Toyota’s challenge in the early days was to produce small quantities of many varieties because of Japanese customer demand. Large batch-building made less sense in this environment, especially when receiving parts from outside suppliers. The assembly plants would be starved for parts sometimes and flooded with parts at other times.
According to Taiichi Ohno, these irregular parts deliveries caused less production at the beginning of a month and lots of production at the end of the month. He called this type of production “dekansho,” after an old song that told of sleeping half the year. These irregularities resulted in lots of waste and made it difficult to produce “small quantities in many varieties.” Toyota’s solution was the idea of level-loading production, or heijunka.
What is level-loading, or heijunka? This is the definition from the Toyotaky.com website… "The overall leveling, in the production schedule, of the volume and variety of items produced in given time periods. Heijunka is a pre-requisite for Just-In-Time delivery.”
The goal of heijunka is to produce the same number of products each time period, generally a day. The result is that daily production volume remains constant. It’s informative that Toyota says that it’s a pre-requisite for Just-In-Time delivery. In other words, heijunka is a requirement. It’s not possible to have a Just-In-Time system without heijunka.
Level-loading (heijunka) is key to reducing waste in production. The lesser-known and emphasized wastes of mura (uneven pace of production) and muri (excessive workload) are countermeasured through heijunka. In addition, excessive inventory and overproduction are reduced.
In order to level production by volume and mix, there must be a detailed production schedule. This schedule should seek to reduce fluctuations in the output rate of each process and the mix of product types within processes. Level-loading satisfies the just-in-time intent of right part, right quantity, and right time.
One requirement to leveling production is to be stingy with the production information. We should avoid providing weekly or monthly schedules to production areas. The production schedule information provided should be only what is required, when it’s required, in the amount required. In one-piece-flow processes this information could be the sequence of items to be built in order (1, 2, 3, 4…).
Many times, these items are built to a buffer, even if it’s a buffer size of one. Buffers within and between processes are necessary for various reasons. Distance between processes, uneven cycle times, eliminate wait time, and machine unreliability are just a few. Additionally, Toyota utilizes buffers to implement heijunka. They decouple inventory which enables heijunka in a downstream department. This is because heijunka requirements are different from department to department.
An example at Toyota is the Paint Selectivity Bank. Here, a buffer is maintained in Paint so that Assembly can pull the correct body type needed regularly to level-load production in Assembly. In fact, this buffer is what enables Assembly the ability to pull. Without this buffer, Assembly would be forced to receive bodies directly from Body and Paint in the order that they’re produced which would cause problems (Mura and Muri) with the heijunka requirements of Assembly. Bodies would come to Assembly randomly (without heijunka), which would lead to unexpected variation in cycle times in various places.
Without heijunka, items tend to be built in batches. This is because each process sees this as the way to be efficient. But this leads to waste. For example, when producing batches of high work content products, the assembly line speed must be set up to reflect this slower assembly time related to the higher work content. The entire production line produces at this pace. Conversely, if the assembly line work pace reflects the low work content products, overall assembly process pace and efficiency are lowered. This is because when low work content products are made at the pace of the high work content products, idle work time results for these processes. It takes less time to produce the low work content products.
There’s another principle that must be adopted to achieve heijunka. This might be a tough one for some. Many times, machines are capable of producing more than required by their customer(s). If utilizing OEE (Overall Equipment Efficiency) as a KPI (Key Productivity Indicator), the process will make products to raise OEE rather than to satisfy system demand. (We get what we measure, right? And why do this anyway? The cost of a machine is already sunk cost. Why create problems to “pay for the machine”?) This can drive overproduction, or waste.
We must do away with OEE as a floor level KPI and, instead, only run equipment when need dictates. This reduces overproduction because some machines are idled for some period of time. If needed, this idle time (found capacity) could be used to produce products for other internal customers. For example, a stamping press at Toyota makes many different parts for several different processes rather than set idle.
Toyota uses Operational Availability (OA) instead of OEE. OA means the machine runs when it’s expected to run. A machine or line may be capable of producing car parts/sub-assemblies at a faster pace than required but only what is required is what is needed. Therefore, the machine is idle for some of the time. At the floor level, the # cars made per hour is recorded against what should have been made per hour. So, if we should make 60 cars/hour (1/minute) in a given hour but actually made 58, we have a discrepancy of two cars, or two minutes of downtime. We record the problem and reason(s) for the downtime and the countermeasure. This provides an hourly/daily/monthly record of our problems and countermeasures which we use for problem solving and improvement activities.
In conclusion, we know that customer orders are random in timing, type, and quantity. But we can smooth out production by implementing heijunka into the production system. Heijunka is key to waste elimination, enables us to pull from upstream, and is a requirement for a Just-In-Time production system.
Phil Ledbetter worked in production leadership at Toyota Motor Manufacturing – Kentucky for 16 years. He’s the author of The Toyota Template, The Plan for Just-In-Time and Culture Change Beyond Lean Tools. Phil is the principal consultant at The Toyota Template.
