Lean production: the end of management whack-a-mole

Michel Baudin, 3/16/99

Everyday, many managers play the factory equivalent of the "Whack-a-mole" arcade game: they undertake focused campaigns to improve one dimension of plant performance without realizing that these campaigns will hurt other dimensions. Other campaigns with a different focus then follow to undo the damage done by the first one, but they usually undo the improvements as well. As soon as the manager's mallet has whacked the "delivery" mole, the "cost" mole rears its head in defiance...

Can we improve any aspect of manufacturing without making something else worse? The answer is yes, and this endeavor is in fact at the heart of lean production. For years, I have been asked for a concise, dictionary-type definition of lean production in 25 words or less. I have resisted. It was always easier to poke holes into others' attempts at such a definition than to come up with my own. Inevitably, they were either too broad, as in: "lean production is the elimination of waste," or too narrow, as in: "lean production is the Kanban system."

The request, however, did not go away. Those many engineers and managers with a classical American or European education can't be satisfied with the idea that lean production is just this multifaceted thing that Taiichi Ohno and his disciples have put together. Their minds cry out for a short statement of what lean production is and how it differs from other approaches. They have worn out my resistance and I will take a crack at it:


Lean production: The pursuit of concurrent improvement in all measures of manufacturing performance by the elimination of waste through projects that change the physical organization of work on the shop floor, logistics and production control throughout the supply chain, and the way human effort is applied in both production and support tasks.

This requires some explaining. Defining lean production as a "pursuit" rather than a system may be the first of many controversial aspects of this definition. Once a system is implemented, it is only subjected to minor tweaking afterwards, but lean production is and will remain a work in progress. Its practitioners, starting with Toyota, are constantly reinventing it. I propose to call it a "pursuit" to emphasize its dynamic nature.

Then, "concurrent improvement in all measures of manufacturing performance" means the end of management whack-a-mole. The definition ties in manufacturing performance to waste and to improvement activities. The relationships the definition intends to express are illustrated in Figure 1 and discussed in detail below. The arrows in Figure 1 highlight examples of these relationships, whereas a complete set would have made the drawing unreadable. The elimination of overproduction, for example, clearly reduces cost, but also improves quality, because it prevents quality problems being buried in WIP, and has no adverse effect on other dimensions of performance.

Figure 1. From improvement projects to manufacturing performance

Likewise, reductions in operator waiting time not only reduce costs but also have other positive side effects, and they have no negative side effects. Operators who are kept active, without being rushed, are less bored and more alert than those who are waiting, they are more likely to react effectively to danger, and therefore safer. In addition, management's unwillingness to waste their time tells them that they are valued and makes them feel more secure in their jobs, which enhances morale.

The types of projects listed on the right side are the means of achieving waste reduction. Walking around the shop floor with a clipboard and writing up cases won't do it. Instead, it is achieved through an array of projects that transform the physical shop floor as well as its information flow and the way people work. The willingness to move and modify machines as needed to facilitate the flow of materials and the movements of people sets lean production apart from many other improvement efforts. Assuming that machines cannot be moved and only addressing information flows, usually by laying a computer system on top of the current floor layout, will not achieve the results of lean production.

Quality, Cost, Delivery, Safety and Morale — collectively referred to as "QCDSM" — comprise the most commonly measured dimensions of manufacturing performance. In any plant, performance in any one of these areas can easily be improved at the expense of the others, as in the following ways:

Managers launch these campaigns for reasons that may be external to the plant, such as demands from corporate management, customer complaints, or competitive pressure. The managers focus on one dimension because they rightly feel that a campaign to improve all dimensions of performance simultaneously would degenerate into little more than motherhood statements and exhortations to be good. Their desire is to act directly on overall performance measures rather than on the system they measure -- that is, the interplay of products, processes, materials, machines and people whose performance is being measured. In other words, they are trying to improve test scores rather than the factory's ability to do what it is being tested on. This is where waste elimination comes in.

The elimination of activities within the plant that produce nothing but waste should not degrade any measure of performance. "Waste" is a common, everyday word used in its common, everyday sense and we should not bother defining it. If needed, we could actually define "waste" as any activity whose elimination would degrade no reasonable measure of performance. In this sense, concurrent improvement in all measures of manufacturing performance is essentially the same pursuit as the elimination of waste.

The most astonishing fact about waste is that there should be any left on late 20th century shop floors. Bring all the production managers to the plant manager's office, tell them they should focus on eliminating waste in their departments, and watch their reactions: "We run a tight ship," says one, "we've been doing it for years." "Talk to the suppliers," says another, "that's where there's money to be saved." "Just get us all the parts on time," says a third one, "and we'll build you everything you need." They perceive the status quo as the proper way to go.

Denial reaches into the pinnacle of academia. Robin Cooper and Robert Kaplan, from the Harvard Business School, are leading researchers in management accounting. In the May-June, 1991 issue of the Harvard Business Review, they wrote: “Industrial engineering over the past 40 years got most of the easy savings from labor, materials, and machine-time efficiencies. Little fat remains to be trimmed from these unit-level activities.” Where in the world did they get that idea?

300 years after Isaac Newton conducted the first time study at the London Mint, 200 years after the industrial revolution, and almost 100 years after Taylor's work and Ford's first assembly line, the typical contemporary shop floor is still full of easy savings opportunities. Observation reveals that half the space is occupied by work in process, while, at any time, 20% to 40% of the people nominally employed as “touch-labor” are in fact waiting for a machine to finish its cycle, waiting for and picking parts, fetching or returning tools, filling out paperwork, or carrying out other activities that do not get the work in process any closer to the shipping dock.

Contrary to what is being taught in schools, the details of operations are in fact designed by shop floor operators with little guidance and no special training, relying on their common sense. Unfortunately, far from being common sense, the methods for eliminating waste from shop-floor operations are counterintuitive. The work, as designed by shop floor operators in traditional plants is almost always wasteful. The operators do not perceive it that way: if they knew of ways to work better, they would. However, being focused on one operation only, the operators do not see the impact of their decisions on the entire process. For example, they like to work in large batches because it is more convenient for them, and they keep producing as long as they have materials so as not to have to explain why they are idle.

They mentally discount the amount of time they spend waiting, walking, fetching, picking and laying out parts and are shocked when confronted with objective measurements from time studies. They underestimate the impact of "a few seconds" here or there. They would be right if they only had to do the job once, but they are not when they must do it a thousand times a day as a link in a chain of several hundred other people. Then it matters enormously, but that fact is not immediately obvious. This is an issue of human nature, not of level of education or native ability: when faced with the need to set up repetitive operations, PhDs make the same mistakes as operators with a High School education.

To make waste easier to identify, Taiichi Ohno classified it in seven categories. The list of 7 major types of waste emerged as a tool to make operators understand and notice waste that was not previously obvious. In addition, it is a finite list, which is reassuring, since anything they do that doesn't fall into one of the seven types is not waste but useful work. The reason it is a list of seven, and not five or nine, is cultural. It originated in Japan, and Japanese lists tend to have seven items, as in "the 7 tools of QC" or "the 7 steps of automation." Lists of ten are more common in the West, and some have designed lists of 10 types of waste that cover the same ground as the Japanese list of 7.

Ohno's seven types of waste are as follows:

While the 7-waste description does throw some light on lean production, it conjures up the inaccurate image of lean production specialists chasing waste around the shop floor with a checklist on a clipboard. This is not the way it works, and this is where the third part of the definition comes into play. Knowing how much gold there is in the mine is one thing; extracting it, another.

The elimination of waste is not undertaken directly but through a series of projects that transform the shop floor and its management.The relationship between these projects, the categories of waste and the dimensions of manufacturing performance is highlighted in Figure.1.

Lean production involves a number of specific improvement projects which reduce or eliminate one or more types of waste. You do not directly attack waiting and overproduction. Instead, you set up cells with one-piece flow, quick setups, etc. and, as a result of these projects, waste in various categories is reduced or eliminated. Most of the implementation time is not spent assessing the amount of waste to be eliminated but identifying and carrying out the improvement projects that work in a particular factory.

It is a key characteristic of lean production that there are no holds barred as to what aspect of the factory's life the projects may affect, ranging from fixture design to organization structure and wage systems. Lean production is a holistic approach, drawing on all the technical, logistical and managerial talent in the plant.And while, in Figure 1, the projects appear as an open-ended laundry list of unrelated items, but they are in fact components of a coherent structure.

Setting up a cell, for example, involves time-and-motion studies that are traditionally viewed as "industrial engineering," but the result may be that, with current process technology, it would need "3.1" operators to run, and therefore 4 people, one of whom would be idle 90% of the time. In this case, the design team challenges process engineers to improve one of the operations in the cell so that it can be run with just 3 operators.

In implementing the cell, maintenance and equipment engineers get involved to prepare the machines to work in an environment with more stringent availability requirements than a job-shop. Once the cell is in place, logistics and production control must make sure that it is supplied with parts and produces at the right pace, and production management must monitor its performance to make sure that it reaches its design targets.

As cells proliferate, they begin to impact the organization structure and human resource management in many ways:

In other words, in lean production, technical changes on the shop floor drive changes in the support and management superstructure.