by Rother, Mike
In those market conditions, there were still some profitable choices available, and thus less need for nurturing products and situations into profitability.
As the need for improvement and evolution became apparent in the mid- to late 1970s, it may have been possible to stay ahead for a while by simply cutting inventories and head count, which might have been bloated. Today, however, we might well be reaching the limits of improving by simply cutting.
The competition was ramping up only slowly, which made it seem as if conditions were not changing all that much.
Interestingly, moving production to lower-cost countries in order to reduce cost—another form of cutting—does not change the underlying system or improve the production process. Some have called this “making waste cheaper,” because it does not actually change the underlying way of doing things.
What Are the Lessons from This History?
Lesson 1
Simply put, after the Model T era the basic attributes of factory flow in the West barely changed during the rest of the twentieth century, as a consequence of the management system. There were, of course, many technological developments since the end of the Model T days, but as Michael Cusumano, in his early 1980s Ph.D. research, and the famous late 1980s IMVP study, both asserted, from 1930 to the 1980s there was little further development in productivity and factory flow (inventory turns) in Western automobile factories. The basic production techniques stayed about the same.
Toyota’s way of moving forward, in contrast, is very much one of adaptation and continuous improvement; of nurturing processes, products, and businesses into profitability by doing what is necessary to achieve target conditions (Figure 4-5).
Figure 4-5. Productivity trends, Toyota and the Detroit Big Three
Source: Michael A. Cusumano, The Japanese Automobile Industry:Technology & Management at Nissan & Toyota (Cambridge, Massachasetts: The Harvard University Press, 1985).
Lesson 2
In the early 1950s the baton of continuous process improvement toward the ideal flow vision was picked up again, and this time by Toyota. In production, for example, Toyota decided to keep working step by step toward something like the early Ford Motor Company vision: a connected and synchronized flow with ever shorter lead time. In fact, both early Ford and Toyota have referred to the production ideal as “one long conveyor.”
Toyota recognized that a main source of low cost is not high machine utilization by itself, but rather when parts flow uninterrupted from one process to the next with little waste in between. For Toyota, striving toward this kind of a synchronized flow meant taking on the challenge of eliminating or reducing the time required to change over between the different items required by the customer.
Lesson 3
The most important lesson to derive from this chapter is that many of us are managing our companies with a logic that originated in the 1920s and 1930s, a logic that might not be appropriate to the situation in which your company finds itself today.
GM’s approach proved highly lucrative during the period of growth and oligopolistic isolation from global competition that extended until the 1970s. It became our model and accepted management practice, and is still taught in business schools today. That means, for many of us the way we currently manage our companies is built on logic that originated in the conditions faced by companies in the U.S. automobile industry during the late 1920s. The problem is not that the logic is old, but that it does not incorporate continuous improvement and adaptation. If our business philosophy and management approach do not include constant adaptiveness and improvement, then companies and their leaders can get stuck in patterns that grow less and less applicable in changing circumstances.
The solution is not to periodically change your management system or to reorganize, but to have a management system that can effectively handle whatever unforeseeable circumstances come your way. The fact that Toyota has maintained much of its same management thinking for the past 60 years is a testament to this. Several of us are curious to see how Toyota’s management system will maneuver and weather the next few decades.
Let us now take a look at that management system.
Notes
1. Keep in mind as you go through this chapter that in retrospect all history is revisionist, and despite my best efforts to dig deep and be impartial, that undoubtedly holds true for this history as well.
2. The elevation drawing is found in: Horace Lucien Arnold and Fay Leone Farote, Ford Methods and the Ford Shops (New York: The Engineering Magazine Company, 1915).
3. Testimony of Edward Gray, Ford Tax Cases, 1927, page 1241.
4. Alfred P. Sloan, Jr., My Years with General Motors (New York: McFadden-Bartell, 1965).
5. Peter Drucker, The Practice of Management (New York: HarperBusiness, 1993). Originally published in 1954.
Part III The Improvement Kata: How Toyota Continuously Improves
Introduction to Part III
In Chapter 2 we saw that the question “What can we do?” often results in scattershot improvement attempts. The more difficult and focused question is, “What do we need to do?”
How does Toyota determine the answer to that question?
Briefly put, the continuously repeating routine of Toyota’s improvement kata goes like this: (1) in consideration of a vision, direction, or target, and (2) with a firsthand grasp of the current condition, (3) a next target condition on the way to the vision is defined. When we then (4) strive to move step by step toward that target condition, we encounter obstacles that define what we need to work on, and from which we learn (Figure P3-1).
Chapters 5 and 6 together comprise a description of the improvement kata. Chapter 5 explains target conditions, and Chapter 6 explains how to go about moving toward a target condition.
Figure P3-1. The improvement kata in brief 75
Although the improvement kata describes a routine for continuous improvement, keep in mind that this kata is also part of Toyota’s way of managing people every day. The psychology of the improvement kata is universal, and at Toyota everyone is taught to operate along the lines of this systematic approach. You will find it applied to many different situations, not just in manufacturing. The content varies, but the approach is the same.
You will also find the improvement kata is practiced at all levels at Toyota, like fractals. The same kata is utilized on both the operative and strategic levels. The scope of the issues addressed with the improvement kata gets broader the higher in the organization you go, but the approach at all levels is basically the same.
The examples in Part III of this book are at the process level in production operations, where I first learned about the improvement kata. The process level is a good place to first focus our attention and learn, since this, along with product development, is where value is added in a manufacturing company. To distinguish between target conditions at the process level and those at higher levels I will sometimes use the phrase “process target condition.”
In production, processes are the individual chain links of a value stream (Figure P3-2), and the word process refers to several different kinds of activity, not just material-conversion activities such as stamping, welding, painting, or assembly. Material handling and scheduling, for example, though not value adding (NVA) in themselves, are nonetheless processes in a production value stream. Such necessary NVA processes should be continuously improved too, in a way that moves the value stream toward the 1×1 flow ideal state.
Figure P3-2. Some examples of processes in a manufacturing value stream
Chapter 5
Planning: Establishing a Target Condition
Once you have experienced the role a target condition plays in Toyota’s improvement kata, you will find it difficult to work without one. You will also discover how difficult it is to explain what a target condition is and its importance to any manager, engineer, or executive who has not experienced it for themselves. That is a Catch–22 we will deal with in Chapter 9.
Over the course of this chapter the target condition idea should become clearer to you, but in the end there is no substitute for learning by doing.
Having a target condition is so important for effective process improvement and management that Toyota will usually not start trying to improve or move forward before a target condition has been defined. This ensures that people’s efforts will be focused on actual needs rather than on various ideas and opinions about what we can do.
A target condition describes a desired future state. It answers questions like:
How should this process operate?
What is the intended normal pattern?
What situation do we want to have in place at a specific point in time in the future?
Where do we want to be next?
Figure 5-1. The role of a target condition
A target condition works like a pair of eyeglasses that helps you focus and see what you need to do. You will discover problems and obstacles any time you establish a target condition and then try to move toward it (Figure 5-1). This is completely normal, and you have two choices:
Avoid the obstacle(s) and move off in a direction other than the vision.
Work through the obstacle(s) by understanding and eliminating its causes.
For example, the employees at the sensor cable company in Chapter 3, who pointed out the problems associated with reducing the lot size from one week to one day, were correct, but what they were pointing out were obstacles, not reasons to change direction.
Seeing Lean Techniques in a New Light
A good way to begin our discussion of target conditions, or target-condition thinking, is to look at some lean techniques we may think we already understand. For each of the four techniques below I will briefly review the technique and then discuss its less apparent but more important intention from the perspective of target conditions.
Takt time
1×1 production (continuous flow)
Heijunka (leveling production)
Kanban (pull systems)
After using these techniques as examples to get us started in understanding the idea of a target condition, I will broaden the discussion to describe important characteristics of target conditions overall.
Takt Time
Takt time is the rate of customer demand for the group, or family, of products produced by one process. Takt time is used most often at assembly-type processes that serve external customers.
Takt time is calculated by dividing the effective operating time of a process (for example, per shift or day) by the quantity of items customers require from the process in that time period (Figures 5-2 and 5-3). “Effective operating time” is the available time minus planned downtimes such as lunches, breaks, team meetings, cleanup, and planned maintenance. Note that unplanned downtimes and changeover times are not subtracted at this point, because they are variables you want to reduce.
Say an assembly process has 26, 100 seconds effective operating time per shift, and over some period of time the customer requires an average of 450 pieces per shift:
The quotient of 58 seconds means that, based on our available time, on average the customer is currently buying one unit every 58 seconds.1
Figure 5-2. The takt time calculation
Figure 5-3. Calculating takt time
How is this number used?
It does not automatically mean you should produce at a rate of one piece every 58 seconds. The actual intended cycle time of an assembly process, called planned cycle time, is usually less (faster) than the takt time. For example, if there is a changeover time between different part types, we have to cycle the process faster than takt in order to compensate for time lost during changeovers. So in a sense takt time represents an ideal repetitive cycle for an assembly process, a cycle at which we would be producing in sync with the customer demand rate—sell one, make one.
The Intention Behind Takt Time
Takt time becomes interesting in our discussion of target conditions when we use it as something to strive for. Two ways to do this are trying to produce consistently to planned cycle time, and trying to move the planned cycle time closer to the takt time.
Trying to produce consistently to planned cycle time means striving to develop a stable process. Many of us track pieces produced per hour or per shift and therefore are unable to answer the question: “At how many seconds per piece should this process be cycling?” We have an aggregate outcome target, but not a target condition, and we get trapped by such outcome metrics because they prevent us from seeing the actual condition of the process. The result is that an astonishing number of processes come close to making their numbers on average, but their output cycles actually fluctuate well out of statistical control from cycle to cycle (Figure 5-4). This condition is not only expensive (requiring extra resources) and adversely affects quality, but many process improvement efforts will simply not stick when the process is out of control.
Figure 5-4. An unstable process
When you have identified the degree of fluctuation from cycle to cycle in a process, the next question becomes: “What should the range of fluctuation be?” With that desired condition in mind, you can then observe the process with an eye to identifying, understanding, and eliminating obstacles to that condition.
Once a process cycles relatively consistently, you have a basis to possibly go further by striving to reduce the gap between takt time and planned cycle time. For example, we might establish a process target condition that includes a planned cycle only 15 percent less (faster) than takt time. As you try to achieve this condition you will again discover obstacles (changeovers, machine downtime, scrap, absenteeism, etc.) you need to work on (Figure 5-5).
Takt time and planned cycle time are only one part of a target condition for a production process, and I am not suggesting that utilizing takt time in this way is the priority improvement for every situation. The point is, we have missed the target-condition intention behind it. Most factories I have visited know about takt time and even calculate it, but so far I’ve found few factories outside of the Toyota group that use takt time as something to strive for in the manner described here. Only then does it become useful.
Figure 5-5. Reducing the gap between takt time and planned cycle time
Note: Toyota subtracts changeover time in calculating the planned cycle time of a process, but not unplanned downtime, which is made up with overtime, as necessary, at the end of each shift, rather than by further speeding up the planned cycle time to compensate for it in advance. This is done to keep problems visible. Of course, to take this approach you need a time gap between shifts that can accommodate such overtime.
I once mentioned to one of Toyota’s supplier support specialists that I’d figured out how to see what one needs to work on at a process. My idea was to ask the supervisor what would happen if we were to slow his process cycle down so it was only 15 percent faster than the customer takt time. The obstacles and objections that the supervisor mentioned would be what we needed to work on!
“Well,” the specialist replied, “the supervisor will be telling you her opinion. To understand the true obstacles, maybe you should build up a little safety stock and temporarily run the process at the slower cycle time. The obstacles that then actually arise are the true ones you need to work on next.”
1×1 Flow
Let us begin by looking at two processes: one without 1×1 flow and one with 1×1 flow. The assembly process depicted below has four workstations, and one operator at each one. There are small quantities of in-process “buffer” inventory between the workstations, as indicated by the inventory triangles (Figure 5-6). The work content each operator has per cycle is shown by the black bars of the operator balance chart.2
Is there a 1×1 flow in this process?
No. Work pieces do not move from one processing step directly to the next. They pass through small buffer inventories.
Is the number of operators correct here?
No. The four oper
ators are not fully loaded up to the planned cycle time. There are extra operators in this process.
What happens if one operator experiences a problem?
Not much. The other operators can keep working because of the buffers between the processing steps.
Figure 5-6. Assembly process with four workstations
Is this process flexible?
Many of us would say yes, this process is flexible, because despite small process problems and stoppages, it can still produce the required quantity every day. With extra operators in the line, the process has the “flexibility” to work around problems and still make the target output.
Now here is the same process, but with the workstations moved a little closer together and the work content distributed in a different manner. There are now two operators who move across the workstations, as shown in Figure 5-7, and no buffers between the processing steps. Takt time and planned cycle time are the same as in the previous diagram.
