Tuesday, April 28, 2009

The A3 and evidence-based healthcare

At the Rona Consulting Group (see "sister sites" at the left), my partners and I help health care organizations transform themselves into quality-conscious institutions capable of making money by employing the principles of the Toyota Management System. Most of the activities we undertake involve teams chartered using Toyota's famous A3 knowledge management system (after which this humble blog is named).

What is an A3?

"A3" refers to a large European paper size (297 x 420 millimeters or 11.69 x 16.54 inches) that is functionally equivalent to the Imperial tabloid (11 x 17 inches). Toyota uses A3 size paper to document its myriad improvement projects and the various team charters and reports generated in the course of project management. These documents are all referred to as "A3s." (For more details, see my book, Hoshin kanri for the lean enterprise (Productivity Press, 2006) or visit my web site (again, see "sister sites" at left) of the same name.)

How does an A3 work?

Essentially, an A3 creates a team charter that is a framework for a collaborative, multidisciplinary, scientific investigation of a problem. Every A3 requires the team leader or "author" to specify a quantitative improvement target and to integrate that target into a succinct narrative that that inspires or compels others to participate in the improvement process. This target statement--and supporting database--provide the framework for gathering and communicating the data that supports analysis and corrective action.

So how is this relevant to health care?

It plunges us into the beating heart of evidence-based health care. Every A3 document rests upon the scientific logic of the Deming Cycle of Plan, Do, Check, and Act (or PDCA). PDCA refers to the planning, doing, checking of a scientific experiment to improve work processes; the "A = acting" bit of PDCA refers to the publication of a new standard based upon a successful, verified experiment.

Tell me again how this is relevant to health care...

So tell me again how this is relevant to health care in some way that is not relevant to other industries? You've got me there, because evidence-based improvement is capable of cutting through the defenses of all types of entrenched interests, including the considerable defenses of management and organized unions universal to industries of all persuasions, including manufacturing and service industries, which of course include the health care industry.

It becomes clear.

Now it becomes clear, at least to the economist in me--I am a trained economist and economic reasoning is a principle unifying thread in this blog--why health care is a "special case," so to speak. Health care is an industry that has the usual, conservative management hierarchy plus multiple unions of considerable power. Actually, I prefer to refer to unions as "guilds." All guilds tend to restrict economic freedom in order to enhance the financial power of their members. Health care's guilds wield very considerable power (for starters, who has not heard of the American Medical Association?), power more potent perhaps (?) than the guild of the United Autoworkers.... In this connection, see Sharon Begley's cheeky Newsweek article, "Why Doctors Hate Science" (http://www.newsweek.com/id/187006).

Moreover, this economically dire situation (dire, that is, from the perspective of a free-market-loving economist) is compounded by the fact that health care organizations tend to have very few competitors, domestic or foreign (although this is changing as the costs of health care rise and the relative costs of travel fall). Predictably, this exacerbates the traditionalist perspective of management and reinforces the monopolistic tendencies of health care's guilds.

At a stroke, we have a logical explanation for a) why health care is relatively (or abnormally) poor with respect to quality and cost, not to mention the ability to adapt to changes in its environment; and b) why the scientific, evidence-based approach of Toyota's A3 knowledge management process is so very very relevant.

An inconvenient truth.

As Jesus so very inconveniently once said, "The truth shall set you free." Toyota's A3 process of PDCA (plan, do, and check your experiment, and then act on the results by publishing new standards) is a rigorous, truth-telling process that any industry may benefit from. But with all of its noncompetitive limitations--entrenched bureaucracy, multiple and powerful guilds, and a pronounced lack of competition--health care will (as surely as night follows day) benefit more than most.

Tom Jackson
Portland, Oregon

Monday, April 27, 2009

Escape from the Panopticon

This post is a continuation of "The paradox of the Panopticon," posted on April 25, 2009...

The way out of the
Panopticon is through organizational redesign and transformation based upon the Toyota Way. In other words, to escape the Panopticon, we must transform organizations into lean enterprises, which have two critical features:
  1. Radical decentralization of decision making; and
  2. Cybernetic control of quality and finance.
Radical decentralization

As every good American should know, and as Jeremy Bentham--architect of the Panopticon--himself taught, the greatest social good is achieved when we let every individual decide for himself or herself how best to conduct his or her affairs. Naturally, to escape the Panopticon, we should want to design organizations that preserve the benefits of free market thinking as much as possible.

The Toyota Motor Company has done something very much like this for large business organizations in the form of kaizen or continuous improvement. At Toyota, it is the job of all employees to find defects and fix them, in as close to real time as possible, without asking for permission. Toyota's model is perfectly general, having been successfully applied in many industries, including healthcare. So long as the scientific method is employed in doing so, employees as well as managers are empowered to implement countermeasures on the spot. In effect, a lean enterprise is a community of scientists conducting and acting upon their own experiments in an internal market for organizational knowledge. In this sense, a lean enterprise is radically decentralized and thus is like that most extreme example of radical decentralization: the free market.

Cybernetic control

With all of those empowered employees running around, however, radical decentralization might degenerate into anarchy without an appropriate control mechanism. Who is to say that employees will not pursue their own experimental interests at the expense of the organization's?

We said that to escape the Panopticon, organizations should be as much like free markets as possible. How are free markets controlled? In a word, they are "cybernetic." The term "cybernetic" is from the Greek for "self-steering." When economists say that they expect markets to "self-correct," they are referring to the cybernetic quality of markets to return to equilibrium without government intervention. When we refer to cybernetically controlled organizations, we are not referring to computerized controls, but to the natural tendency of appropriately designed organizations to "self-correct" without executive intervention.

This is a far cry from management accounting, the control mechanism employed by most large organizations in the world today. Management accounting mirrors the Panopticon and Foucault's Discipline and Punish more readily than the laissez faire of the marketplace.

Within a lean enterprise, hoshin kanri is the mechanism that provides the coordinating benefits of organizational control--including financial control--without compromising the utilitarian ideal of freedom by instilling in every employee a common notion of the organization's strategy. Through an elaborate process of negotiation called catchball, all managers within the lean enterprise establish both the financial targets and the process improvements by which those targets will be met. For a more detailed explanation, please see my earlier post, "What is lean management?" The result of this process is a system that is essentially cybernetic.

What does this mean?

Lean organizations like Toyota are able to produce goods and services at much higher quality and far lower per unit costs and with far fewer layers in their management hierarchies than Panopticons like General Motors:
  • Womak and Jones (in their groundbreaking The Machine That Changed the World) estimated that a lean enterprise produced twice the product in half the time at half the normal defect rate with a fraction of the required work-in-process inventory; moreover,
  • lean enterprises require fewer managers (at its height, GM had as many as 33 layers of management, while Toyota has about 9.)
Although no one is saying that Toyota is a free market, it is clear that the radical decentralization of empowerment and the cybernetic control of hoshin kanri create an extraordinarily productive and adaptive organization form. My point is that in doing so Toyota has resolved the Prinsoners' Dilemma and shown the way out of the Panopticon.

Tom Jackson
Portland, Oregon

Saturday, April 25, 2009

The paradox of the Panopticon

In his book, Discipline and Punish (1975), French philosopher Michel Foucault characterized modern organization as a Panopticon, a special type of prison. The brainchild of British utilitarian, Jeremy Bentham, the design of the Panopticon gives its Warden complete control over his Prisoners through a clever combination of spatial organization and mirrors. From his exalted place at the center, the Warden can see the activities of all the Prisoners in all places at all times. Of course, it is not possible for the poor Prisoners to see when they are being observed by the Warden. (Bentham's blueprint for the Panopticon appears below.)

Foucault's analogy is perfect. From French architect Le Corbusier's "machines for living," to the office cubicle, to the security camera, to the monitoring of employees' internet traffic, to airport X-ray machines, to RFID implants, to the mind-control technology of Steven Speilberg's Minority Report, the Panopticon resonates with our own experience and expectations of Big Business, Big Government, and George Orwell's Big Brother. Of course, 9/11 has lately redoubled Big Brother's diligence and intensified the claustrophobia and paranoia of the Prisoners.

The Panopticon is a trap, of course, for the Warden as well as the Prisoners. For while the Warden may observe what the Prisoners are doing, the Warden may never know what the Prisoners are experiencing. Or thinking. Not even coercive interrogation can help. All of which impoverishes the poor Warden, intellectually and socially as well as economically, as game theorists have explained for over half a century in a famous example known as the Prisoners' Dilemma. Economist Oscar Morgenstern and mathematician John von Neumann showed in their groundbreaking Theory of Games and Economic Behavior (1948), that, assuming that Prisoners are forbidden to coordinate their responses to the Warden's questions, the result of interrogation will be to limit social utility by denying shortened jail terms or even freedom to either or both Prisoners. I should hasten to add that there is an implicit assumption in the Prisoner's Dilemma that the Warden's compensation package turns not upon truth-based convictions, but upon convictions at any cost. In other words, the Warden is not rewarded for punishing the right Prisoner, but for punishing any Prisoner. The Warden can only interrogate Prisoners imprisoned in the Panopticon. Who knows whether or not the Warden has imprisoned the Guilty One?

The Prisoners' Dilemma is a potent criticism of Bentham's utilitarian philosophy, which holds that individuals pursuing their own interests naturally creates the greatest utility for society as a whole. If we accept Foucault's concept of the Panopticon is a correct characterization, then we must also conclude that modern organizations are inherently stupid, for clearly the Panopticon is not structured for learning.

In future posts I will explain how we may all--Warden as well as Prisoners--escape the Panopticon through better organizational design.

Tom Jackson
Portland, Oregon

Friday, April 24, 2009

Gemba v. Maya

Lean management consultants who specialize in the Toyota Way always say, "Go to Gemba." Gemba is the real place, the here and now, where life actually happens. The reason that we must go to Gemba is because, if--as managers--we rely upon computerized databases or abstract reports, we can never be sure that we actually know what is happening. Regardless of the millions, or hundreds of millions of dollars lavished upon information technology, the information contained by systems of hardware and software is by definition out of date--by at least the speed of light, and frequently much, much longer. The world of symbolic knowledge, the world of words, numbers, bits, and bytes is inherently incomplete. Symbols after all can only signify. They cannot also be the real things to which they refer. At least they cannot be the only things; otherwise language and numbers are pointless.

So, we must go to Gemba and make direct observations for ourselves in, as we say, real time.

The distinction between Gemba and Maya, the world of symbolic knowledge, is rather a Buddhist idea. Buddhist philosophy distinguishes between the absolute reality of Sunyata, nothingness, or the void, on the one hand, and the relative reality of Maya, on the other hand. Maya is often translated as "illusion." The Sanskrit root of the word Maya means to measure. Thus, Maya is measured reality, that is, the reality of words and numbers that organize the onslaught of experience into things separated in spacetime and distinguished by differences such as form and color and number.

Like good Buddhists, lean management consultants do not deny the practical importance of Maya, that is, of databases and reports. The abstract information contained in databases and reports is, however, ultimately untrue when compared to Gemba. And so to Gemba we must go, to refresh our knowledge of what is real.

Why bother with Buddhist philosophy on a blog about lean management? Because, it takes us deep into to the heart of Toyota's Way. If, as managers, we must go to Gemba to know the real, what happens when we return to Maya, as we must? Surely things will fall apart when we take our eye off the ball, so to speak. We will lose control. That is, unless we rely upon our people to be in Gemba themselves. In other words, we must decentralize decision-making, radically, to ensure that decisions can be made in Gemba, in real time, based upon direct observation of the process. So it is that Toyota's production workers or nurses at the Virginia Mason Medical Center in Seattle can "stop the line" whenever a process defect occurs. No one else is there to observe the problem; and no one else can take more rapid action to prevent the problem from causing more harm.

The need to decentralize in order to control places management in a strange new role. We are of a piece. We are a system. Thus to function better as a system, we must raise the level of all participants in the system. We must train everyone to go to Gemba, to "be here now," to distinguish between what is real and what is not real. As the Buddhists say, "None of us is free until all of us are free."

Tom Jackson
Martinez, California

Thursday, April 23, 2009

The Matrix and the Balanced Scorecard

Toyota, more famous for its production system than its organizational structure, has operated a highly successful matrix or "cross-functional" organization since the early 1960s. Theoretically, matrix organizations should be better at solving complex problems requiring a multidisciplinary approach. Toyota uses its matrix to routinely reduce its operating costs as well as solve quality problems. Royal Dutch Shell famously abandoned its matrix structure in the 1990s, presumably to cut costs. Why should Toyota find the matrix tenable while Shell did not?

The answer may be the "balanced scorecard," or what in Japan is known as "hoshin kanri." (Actually, the balanced scorecard is a much-watered-down version of hoshin kanri, of which more in future posts.) Japanese authors claim that while Japanese companies experimented with cross-functional or "matrix" organizational structures, it was not until the adoption of hoshin kanri that they were able to make the matrix work. Why might this be true?

The answer to this question lies in transaction cost economics. Transaction cost economics is based upon the idea that the fundamental unit of organizational analysis is the contract. Organizations arise because management hierarchies, by creating standards of communication and behavior, offer a more cost effective way than free markets of concluding and enforcing the various contracts needed to do business (management contracts, labor contracts, contracts with suppliers, contracts with customers). If human beings were infinitely intelligent and markets perfectly efficient, we could conclude and enforce the same contracts in less time and at lower cost through the one-on-one bargaining of the marketplace. Actually, if the assumptions of traditional microeconomics were true, markets would clear in an instant. But, alas, human beings have small, slow brains, and markets periodically melt down, as we know all too well. Thus the management hierarchies of business frequently offer a quicker, cheaper way to get business done (i.e., conclude and enforce all those contracts).

So what does this have to do with matrix organizations and hoshin kanri?

Well, the matrix is complicated. Some people believe that although it promises to make organizations good at solving problems, it isn't worth the effort. A common complaint is that within a matrix organization I have more than one boss. It is not clear who is in charge and thus my reporting relationships present a control problem. (One of my clients, once a player on Wall Street, was fearful that the matrix violated Sarbanes Oxley.) Implicit in this argument, however, is an assumption that the cost of an effective cross-functional agreement between my two or three bosses as to my priorities outweighs the benefits of my being on a cross-functional team. In other words, in practice the matrix is so expensive to control, that it is better to resort to the traditional command-and-control structure, even though command-and-control isn't very good at solving complex problems. More simply, the cost of solving complex problems in a command-and-control stucture is less than the cost of controlling a matrix organization.

This is where hoshin kanri comes into play. Hoshin kanri greatly reduces the transaction costs of concluding and enforcing agreements and thus enables more complicated, and more effective matrix structures. At its core, hoshin kanri is all about contracts or, as we know them, team charters, or, more recently, as A3s (hence, The A3 Post). The hoshin system is essentially a system of A3s (the contracts) that, through a painstaking process of negotiation known as "catchball," completely specifies the operative relationships between all managers, both vertically (top down/bottom up) and horizontally (cross-functional). At Toyota, for example, even supervisors and suppliers are included in this process of specifying the organization's structure.

In addition to creating a completely specified matrix structure, hoshin kanri produces contracts that are less costly to enforce. Although the catchball or negotiation process is not perfectly voluntary, it has many of the same features as a free market negotiation. Let's say that I am a supervisor negotiating my A3 contract with my manager. My manager tells me "what" to do by deploys a target to me that is pretty much nonnegotiable. It is an offer I can't refuse. In response, however, I am permitted--in fact I am encouraged--to recommend "how" I plan to achieve that target and to specify the resources I will require to do so. We might say that the A3 contract is "quasi-voluntary" because it results in a meeting of the minds. By leaving the negotiation open ended, my manager gets me to "buy in" to the agreement.

Now we can see how hoshin kanri enables matrix organizations. Matrix organizations that do not employ hoshin kanri risk creating organizational structures that are not completely specified and, therefore, in which important relationships are open to interpretation by managers with small slow brains. Mischief will ensure, because, given a choice, a choice created by lack of clarity, managers will pursue their own agendas rather than the organization's. Also, the unclear relationships are usually the cross-functional ones that would help us solve complex problems, because the traditional management hierarchy makes top down/bottom up relationships crystal clear. In addition, matrix organizations in which command-and-control managers mandate both "what" and "how" risk creating involuntary or, shall we say, "quasi-inconscienable" contracts that are, predictably expensive to enforce, because there is no buy-in. So, without hoshin kanri, a matrix organization inflicts a double whammy upon itself. First, the incomplete specification of the contracts that define the organization leaves too much to guesswork; organizational focus is lost. Second, the involuntary nature of command-and-control contracts makes it less likely that subordinates will adhere to the contracts voluntarily, thus creating the need for extra controls. It's no wonder Shell Oil abandoned the matrix! Much better to be a dull, plodding command-and-control organization....

Meanwhile, Toyota's ascendancy demonstrates that, with the right management control system--viz., hoshin kanri, the matrix organization can deliver on its promise of flexibility and adaptability in practice as well as theory.

Tom Jackson
Martinez, California
April 23, 2009

Wednesday, April 22, 2009

A small technicality

To my readers, who may have noticed a lapse in my blogging, the fault is not all mine. Although I have been an avid Apple fan ever since I bought a Mac SE 30 and matching portrait monitor in 1988, I was recently disappointed in Apple by the odd failure of its otherwise wonderful iWeb software. iWeb's blogging program has a bug that deposits unwanted text above the blog header--like a fly on the windshield--of every blog post. Unfortunately, the text cannot be removed or even edited.

Enough about Apple. Rather than clean the windshield, which as I explained was not possible, I bought a new car, from Google. I have ported all previous posts and redirected internet traffic to the new blogspot URL.

With my apologies,

Tom Jackson
Martinez, California
April 22, 2009

The 5 rules of lean DNA

Steven Spear and H. Kent Bowen did an admirable job of describing what makes Toyota tick in their Shingo Prize-winning article, “Decoding the DNA of the Toyota Production System,” Harvard Business Review, Sept.-Oct., 1999. In that article they defined four “rules” that described how Toyota’s production workers carry our their activities (Rule 1), make connections between customers and suppliers (Rule 2), create pathways of material and information flow (Rule 3), and make improvements (Rule 4). In this post, I will argue in favor of a 5th Rule of lean DNA that describes how Toyota develops leaders. See my short article, The 5 rules of lean DNA. You will find Spear and Bowen’s first four rules restated there as well, with minor modifications.

It is often said that the Toyota Production System is a philosophy. According to the Merriam-Webster online dictionary, a philosophy is a theory underlying or regarding a sphere of activity or thought, or the most basic beliefs, concepts, and attitudes of an individual or group. This does not seem to be particularly helpful, particularly because Toyota is so very practical. I prefer to think of the Toyota Production System as a culture. The concept of “philosophy” is too small to contain Toyota’s system.

While definitions of culture abound, I prefer to think of culture in four dimensions,

A set of ideas (concepts, values, etc.);
A set of tools (i.e., a technology, including methods of communicating, i.e., language, and methods of getting things done, e.g., agriculture, manufacturing, etc.);
A method of adapting to changes in the environment; and
A method of transmitting all four of these elements to future generations.

Perhaps it is such a notion of culture that Toyota has in mind when it speaks of The Toyota Way. See Jeffrey Liker’s The Toyota Way. I cannot claim to be an anthropologist, but my definition of culture bears some resemblance to anthropological definitions of culture.

Toyota’s ideas

Toyota’s principal ideas may be summarized as follows:

Standard work. Work should be standardized as to task, sequence, time (i.e., takt time), and work-in-process inventory;
Just in time. Material and information should flow without interruption (i.e., one-piece flow) to meet actual customer demand; where you can’t flow, pull material and information according to average expected customer demand.
Jidoka or autonomation. Human work (the work of observation and thinking) should be separate from the work of unconscious machines and processes. All interfaces between--between process customers and process suppliers, and between employees and machines--should be perfectly unambiguous so the defects and/or equipment problems can be discovered and fixed in as close to real time as possible.

Those familiar with the Toyota “house” diagram will recognize these three things immediately. They are also consistent with the first three “rules” of lean “DNA” as articulated by Steven Spear and H. Kent Bowen in their Shingo Prize-winning article, “Decoding the DNA of the Toyota Production System,” Harvard Business Review, Sep-Oct, 1999.

Toyota’s tools

Here is a famous short list of lean tools, drawn from Hiroyuki Hirano’s JIT Implementation Manual (Productivity Press):

Standard operations (standard work)
Flow production (just in time)
Changeover (quick setup)
Visual control
leveling (level loading or heijunka)
quality assurance (successive checks, self checks, and poka yoke)
maintenance and safety (total productive maintenance)

Of course, we would need to create additional lists for different business functions such as engineering, marketing, and so forth. Generally speaking, Toyota’s tools can be identified by their incorporation of the Deming cycle. In other words, regardless of whether it is a tool of manufacturing, of service, or of administrating, a lean tool is a tool designed to adapt to change. See below.

Toyota’s method of adapting to change

As far as I can tell, Toyota has two basic methods of adapting to change:

The Deming cycle, or PDCA, with stands for Planning, Doing, Checking, and Acting upon a scientific experiment to continuously improve upon standard work.
Hoshin kanri, or strategy deployment (also known as strategy deployment). Hoshin is the systematic application of the Deming cycle to the articulation, deployment, and execution of a company’s long term strategic plan. Please see my book, Hoshin kanri for the lean enterprise (Productivity Press, 2006).

The first method listed here, the Deming cycle, is consistent with Spear and Bowen’s four “rule” of lean “DNA.” Their article was based upon observations made on Toyota’s shop floor and did not necessarily support generalizing to management systems such as hoshin kanri. It is clear, however, that since the early 1960s, Toyota has used hoshin kanri to develop its organization. In particular, hoshin kanri has been a critical part the cross-functional management of its value streams, as well as the interorganizational management of its supply chains.

Toyota’s methods of transmitting its culture to future generations

Toyota has essentially two basic methods of transmitting its culture to future generations:

The A3 process, as described by John Shook in his recent Shingo Prize-winning book, Managing to Learn (Lean Enterprise Institute, 2008). The A3 is a charter document printed on large-format paper (in America, we would print on 11” x 17” paper). It is essentially a contract between one manager and another by means of which the first manager delegates responsibility for developing countermeasures to resolve a problem. The method is significant as a leadership development method because it places senior managers in the role of coach or, in the Japanese style, sensei (teacher). To see just how seriously different this method can be from American “take charge” management, see Steve Spear’s “Learning to Lead at Toyota,” Harvard Business Review, May 2004. It is important to note that the A3 process is actually a feature of hoshin kanri (see above) as it is practiced at Toyota. The hoshin process ensures that all senior managers participate as coaches to their direct reports. See my book, Hoshin kanri for the lean enterprise (Productivity Press, 2006).
Systematic training in the ideas, tools, and methods of adaptation listed above. The style of training has been heavily influenced by the U.S. Government’s WWII program Training within Industry, which introduced many of the concepts and practices we recognize as “lean” today. Toyota’s Supplier Support Center, which for many years provided free training to its promising suppliers, is more evidence of the strength of Toyota’s belief in the power of improving human resources through training.

The A3 process is consistent with Spear and Bowen’s 4th rule insofar as they refer to the need to have a coach to oversee the use of the scientific method (PDCA) by employees on the shop floor. The A3 process, however, does not extend that far. As part of hoshin kanri, the A3 is for managers only.

Tom Jackson
Portland, Oregon
April 5, 2009

Who needs accountants?

H. Thomas Johnson writes in To Become Lean, Shed Accounting, that Toyota does not allow accountants on the shop floor. It is no wonder. What would be the purpose? Everything is already under control. Accountants on the shop floor would be, well, muda or waste.

To be more specific, Toyota’s standard work controls six basic things:

exactly how each task in a process is performed,
the sequence of such tasks;
takt time (i.e., the time in which tasks must be performed to meet demand);
the work-in-process (WIP) inventory required to meet takt time;
the documentation of standard task, sequence, takt time, and standard WIP by front-line employees; and
the real-time auditing of standard work by the very people who create it.

My team of experts and instructional designers at Productivity Inc. built the first five of these elements into the Ford Motor Company’s first companywide training in standard work in 1996. The sixth element may come as a surprise. But read on.

Standard work, as defined by Toyota, integrates quality (defined by elements 1. and 2., task and sequence) on the one hand, and cost (defined by elements 3. and 4., takt time and work-in-process inventory) on the other hand. Through the rigorous application of successive quality checks, self-checks, mistake-proofing (all on the foundation of 5S and visual control), front-line employees themselves audit not only the quality performance but also the financial performance of the organization. Consequently, Toyota does not need inspectors at the end its production lines, nor does it need accountants at the end of the month to tell it whether or not it made money.

As opposed to the periodic audits of management accounting, Toyota’s employees perform what are in effect real-time audits. What could be more effective? Well, nothing. Unless you want to discover and fix problems on a monthly basis as opposed to discovering and fixing them right now.

Tom Jackson
Portland, Oregon
April 4, 2009

The cybernetic control of flow

Why is takt time so easy on the one hand and yet so difficult on the other. Everyone understands the simple concepts of the seven deadly wastes. Everyone understands the simple idea of working to the “beat” of the market. Few people have difficulty drawing a current state value stream map. But ask people to understand FIFO lande, supermarkets, and heijunka? What has your experience been?

I believe that the difficulty lies in the failure to teach the idea that a lean organization is cybernetic, that is, self-controlling or self-regulating. Truly, lean enterprise means the elimination of schedules as we know them.

In constructing a lean future state map, there are two major questions:
Will the organization be entirely scheduled by customer demand? (I call this “no hands” scheduling.”)
Will the organization have its own internal schedule or pacemaker?
If the organization is designed to respond entirely to customer demand, then the market is the pacemaker. We may pull information or material from the receiving dock to the shipping dock using one-piece-flow and supermarkets. If the organization is designed to have an internal pacemaker, then we may pull material from the shipping dock to the pacemaker, and flow material from the pacemaker to the shipping dock and from there to the customer.

Sounds easy, right? The trick is to help people see the progressive auto-mation or de-scheduling of decisions about the sequence and timing of information and material movement. Clearly, on the shipping dock, in a just-in-time system every movement should be dictated by customer demand. That is, the customer signals its requirements to us, and we ship. The question in lean scheduling is: To what extent may we link the rest of the supply chain to customer demand without interposing moderating buffers of inventories in FIFO lanes and supermarkets.

FIFO lanes give us the least amount of insulation from customer demand, because once an item of information or material enters a FIFO land we cannot change its sequence and we can only change its timing to a limited extent. There is no queue-jumping in a FIFO lane. Everything else in the system had better be moving in the same sequence and at the same pace, because there will be no more signals from the internal pacemaker, only the customer. If anything is out of sequence or moving at the wrong speed, there will be mismatches when subprocess intersect.

Supermarkets give us the most insulation from customer demand, because once an item enters a supermarket, we can still change either its sequence or its timing. Queue-jumping is allowed when the pacemaker says so. And, depending upon the circumstances, I could sit on the shelf for a relatively long time before the pacemaker asked me to move on.

In sum, we may want to think of the pacemaker as a traffic cop whose job it is to interpret the customer’s demand signals, which are insufficiently clear to achieve information or material flow. Material in supermarkets is like cars in a major parking lot next to the highway of flow. The pacemaker receives a certain amount of information from the customer in real time that permits him to choose who should go first and should go second up the ramp to the highway. It doesn’t matter where or when you parked. Even if you were the last car driven on the lot and farthest from the exit, you may be asked to exit now.

Once you’re on the highway, however, the rules change. On the highway of flow, there is only one lane, and no speeding, tailgating, or passing.

Tom Jackson
Portland, Oregon
April 1, 2009

Forget about material flow

Most lean consultants will tell you that lean enterprise is about three things: material flow; material flow; material flow. And this makes sense, that is, until you think about it. As we know, lean manufacturing techniques have been successfully applied to many nonmanufacturing processes, most recently to healthcare. So obviously whatever lean is about, it is about information as well as material flow.

In fact, it is all about information flow. Think of it this way. Organizations are great big computers. In other words, organizations are problem solving machines. Consider healthcare. A patient visits the doctor. The doctor gathers information about the patient by reading the patient’s history, looking at the patient, taking her temperature, etc. Then the doctor processes the information, looking for patterns that will indicate the patient’s condition. Once the condition is known, the doctor prescribes and sometimes delivers a treatment.

According to economist Oliver Williamson, organizational structure is not necessarily what we think it is, namely, an organizational hierarchy or the degree to which our supply chain is integrated vertically or horizontally. It has to do with we structure the decision-making process within the organization and how well manage our organizational knowledge. According to Williamson, two things determine the effectiveness of organizational structure. The first is the degree to which decision-making is decentralized. Decentralized decision-makers can find and fix defects in the process faster than one centralized decision-maker. Decentralized decision-making also leaves more time for leaders to think about the future. Unfortunately, being only human, we tend to pursue our own interests instead of the interests of the organization to which we belong. So, the second thing is the effectiveness of organizational control. Organizational systems that can balance decentralization with control achieve an organizational state of grace known as “double-loop” learning. In other words, the organizations learn on two levels. According to Chris Argyris and Donald Schon,

When [an] error detected and corrected permits the organization to carry on its present policies or achieve its presents objectives, then that error-and-correction process is single-loop learning. Single-loop learning is like a thermostat that learns when it is too hot of too cold and turns the heat on or off. The thermostat can perform this task because it can receive information (the temperature of the room) and take corrective action. Double-loop learning occurs when error is detected and corrected in ways that involve the modification of an organization’s underlying norms, policies and objectives.See Argyris and Schon, (1978) Organizational learning: A theory of action perspective, Reading, Mass: Addison Wesley, 1978, pp. 2-3.

In the history of business, General Motors is known to have made a breakthrough in organizational learning when it decentralized its decision-making into divisions (Chevrolet, Buick, Cadillac, etc.). At the same time, GM imposed a new system of financial targets and internal audits to control its divisional presidents. (GM’s control system was in fact the precursor of management accounting, which is used to control most organizations today.) In less than a decade of creating this new structure, GM had overtaken the Ford Motor Company, which never recovered its former position as the world’s leading automaker. GM beat Ford for two basic reasons: 1) GM could learn what its customers wanted and provide it faster than Ford; and 2) GM’s leaders had more time to think about the future and could therefore adjust their “norms, policies, and objectives” more readily than Ford. In other words, Ford was a “single-loop” organization; GM was “double-loop.”

With the concept of double-loop learning in mind, we can now understand Toyota’s eclipse of GM, Ford, and most of the world’s automakers. But in order to do so we have to forget what we know about material flow and focus instead on organizational structure and organizational learning. When an economist looks at Toyota, two innovations stand out. First, Toyota has decentralized its decision-making, and done so radically. Individual employees on the front line of operations are empowered to stop the production line, if necessary, to prevent defects from passing to the next process. The second thing is Toyota’s control system, which is not GM’s management accounting. According to H. Thomas Johnson, Toyota does not permit accountants on the shop floor! In place of accounting, Toyota installed hoshin kanri or strategy deployment, a method that aligns managers to strategic intent through a careful process of negotiation known as catchball. See www.hoshinkanri.biz for additional information about this method. Hoshin kanri creates and reinforces double-loop learning in four ways.

By asking managers to focus on a few, critical objectives, it encourages them to delegate, thus decentralizing decision making.
It provides a framework for feedback and control through regular, frequent, short meetings;
It provides a framework for coaching and mentoring direct reports so that they will become effective problem solvers who also know when and how to decentralize decision making;
Reflection on the companies norms, policies, and objectives is an explicit step in the process.

Tom Jackson
Portland, Oregon
March 29, 2009

What is lean management

Lean management boils down to two innovations in organizational structure and management control:

The radical decentralization of decision making; and
A cybernetic financial control system, known as profit management.

Radical decentralization

Radical decentralization is the empowerment of the entire workforce to find defects and fix them, and the quicker the better! It is most simply described as the combination of standard work and continuous improvement. Standard work requires every employee to perform each task in the same way and in the same sequence. Continuous improvement requires every employee to participate in changing standard work, when necessary, using the scientific method. As a matter of history, standard work and continuous improvement are part of total quality management (TQM), which Toyota implemented during the period 1960-63. For a more complicated but more elegant explanation of how Toyota dealt with standard work, see Spear and Bowen, “Decoding the DNA of the Toyota Production System,” Harvard Business Review, Sept.-Oct., 1999. (Toyota actually beefed up standard work by requiring employees to perform standard work sequences within a standard time and with a standard amount of work-in-process inventory. We will return to this point when we describe profit management.)

Profit management

Profit management is a highly refined system of widely deployed financial targets and real-time auditing on the front line of operations. Profit management hinges upon a technique of strategy formation and execution known as hoshin kanri or policy deployment, which in the late 1950s became an integral part of the peculiar Japanese version of TQM. Essentially, hoshin kanri is a knowledge management process that makes systemwide adjustments to standard work. Whereas continuous improvement aims at small, incremental improvements, hoshin kanri aims at strategic change in quantum leaps. Hoshin kanri is the ne plus ultra of strategic change management. Moreover, companies that practice it, really practice it, are constantly adapting to their increasingly chaotic competitive environments. See Cooper and Schlagmulder, Target Costing (Productivity Press).

Originally, hoshin kanri was a quality methodology. At Toyota it became much more. This may have been the result of Toyota’s unique interpretation of standard work. In its simple form, standard work requires employees to perform each task in the same way and each sequence of tasks in the same order. This will certainly reduce process variation, no matter how long you take to get the job done. No doubt inspired by its just-in-time approach to production management, Toyota added two dimensions to TQM’s standard work: time and WIP (work in process inventory). At Toyota, employees must not only perform standardized sequences of standardized tasks, they must do so in a standardized amount of time (takt time) and with a standardized amount of WIP. By incorporating takt time and standardized WIP into its definition of standard work, Toyota integrated its financial and quality systems. For, if you can control task, sequence, time, and WIP, you control cost as well as quality (assuming your standards are reliable).

By adopting hoshin kanri, Toyota made its new, integrated system fully dynamic. Hoshin requires an organization to deploy, which is to say communicate, challenging improvement targets to every one of its managers (top, middle, and front line). At first, such targets were limited mainly to improvements in quality. Toyota expanded the system to include other types of targets, including cost reduction targets. Thus was born “kaizen costing,” a companywide effort that placed Toyota on a path of systematic cost reduction as surely as stepping on the “down” escalator at the department store. (At the same time, Toyota developed the well-know but poorly understood approach to new product development known as “target costing.” Neither kaizen nor target costing are stand-alone inventions, but are part of the broader system of profit management. Again, see Cooper and Slagmulder.)

Radical decentralization, revisited: real-time audits

To avoid overwhelming the reader, I have postponed the discussion of a critical element of radical decentralization, namely, real-time audits. Toyota super-expert, H. Thomas Johnson, reports that Toyota does not permit accountants on its shop floor. How can that be? Well, because of standard work, everything on the shop floor is well under control, and accountants concentrate on external reporting.

Earlier I mentioned Toyota’s zeal in adopting TQM. My evidence is Toyota’s justly famous “andon” or “alert” system, which places the power to stop an entire factory in the hands of every employee. At Toyota, an “andon” cord (sometimes it is a button) literally hangs over the heads of the workers. There are strict rules about when to pull it, but everyone knows that when conditions deteriorate to a predetermined degree, pulling the cord is part of the job. Managers come running to help solve the problem and get things running again. You know this already. What you don’t know is that this is not merely a quality control; it is a cost control, too. Toyota’s andon cord amounts to real-time auditing.

At Toyota, a defect is more than a dented fender, it is anything that diminishes customer value. Defects include material defects and errors in standard work. To grasp this fully, recall the elements of standard work, Toyota style: task, sequence, takt time, and WIP. Of course, a dented fender is a defect. Trying to put it on upside down is tantamount to a defect. If a sequence of tasks cannot be performed within takt time--that’s a defect, too. If employees allow WIP to build up (or fall below) strictly defined limits--even that’s a defect. And, if necessary, the andon cord is pulled and managers come running.

The modern corporation is dead

With the twin innovations of radical decentralization and profit management, Toyota has reinvented the modern corporation. Ironically, Toyota’s most prominent victim, General Motors (GM), reinvented the modern corporation almost 90 years ago with a similar pair of innovations. GM decentralized decision-making by creating corporate divisions (think of Chevrolet, Cadillac, Buick, etc.). Then, to control the powerful presidents of its new divisions, GM imposed a new system of financial targets and internal audits, a system that has evolved into present day management accounting. GM’s innovations were so powerful that they almost put Ford out of business in 1927.

Do we see a pattern? Toyota has decentralized anew and imposed a new system of control. This put GM on the ropes long before the our current economic crisis.

Tom Jackson
Portland, Oregon
March 16, 2009

Too big to fail?

The strange connection between lean enterprise and antitrust law
Simon Johnson, a rising star in the world of recession/depression commentators, stated on a recent appearance on Terry Gross’s essential NPR (National Public Radio) program, Fresh Air, that the current economic meltdown warrants a revision of U.S. antitrust law. See Johnson’s lively blog, Baseline Scenario. This caught my attention. Mssr. Johnson is furious (as am I) that banks, especially—or, for that matter, any companies (GM, for example)—should become “too big to fail.” The reason is that if they become too big to fail they can hold taxpayers for ransom—a ransom to be paid by the taxpayers (that’s you and me, incidentally).
So, what does lean have to do with antitrust? Of course, this is a blog about lean enterprise. Well, a funny thing happened on the journey to lean enterprise. When we think of “lean enterprise” we naturally think of the giant Toyota Motor Company. One reason why people understand the phrase “lean enterprise” is because Toyota is so very big. But there is an interesting twist in this plot: Despite Toyota’s bigness, the famous Toyota Production System may obsolete the much-venerated idea of economies of scale.
Run for the doorsills and brace yourself for a 9.0 earthquake. Savvy lean thinkers already know that you don’t have to be big to be lean. As I often say, give me a process with more than one step in it, and I can make it run faster, better, cheaper—without major capital investment. But how can this be? We Americans, and, to a more or less equal degree the rest of the world, have been brought up on the idea of economies of scale. Bigger, scalable technology is better. Right? Of course. Because, if we “scale” our technology, we can get more throughput and therefore more revenue at an ever decreasing per unit cost. Everything in the consciousness of the global businessman is based upon “bigger is better” and the “scaling” of technology. Ergo, economies of scale.
Not so long ago, we believed that economies of scale were purely knock-your-head-on-the-floor physical—and there is of course some important physics to support this notion. Later, however, we learned that economies of scale could be justified for other reasons as well, including the phenomenon of organizational learning. This was the essential insight behind the experience curve. The experience curve revealed that every time--by whatever means, animal, vegetable, or mineral--you doubled output, you reduced cost by a “scale invariant” percentage that ranged from 5% per unit in the case of very automated processes to 30% or more for very non-automated processes such as pre-computerized administration. Thus, the “bigger” the operation, the lower the cost, and therefore the better for consumers. This was symptomatic of the logic of much of the Reagan Era deregulation that led to the creation of businesses “too big to fail.”
The logic of lean enterprise inverts the consumer-oriented cost reduction logic of Big Business. It is no longer necessary to be big to be cheap. While economies of scale depended mainly upon investment in large machines and large factories and billions of lines of computer code that pump parts, services, and bit of information out faster and faster to meet expected demand, the economies of lean enterprise depend upon investments in individual workers who consciously pace and re-pace the production of parts, services, and information flows to meet actual demand. Very often, the adoption of lean production requires companies to slow down, rather than speed up. To be cheap, one does not need to be big, one merely has to be quick. The Boston Consulting Group’s renowned strategist George Stalk makes similar argument for “sidestepping economies of scale” in his new book, Five Future Strategies That You Need Right Now (Harvard Business Press).
What are the implications for public policy? The relaxation of antitrust law over the last thirty years has been premised principally upon the notions of economies of scale and the efficiency and reliability (until recently) of free markets. Bigger is better for the consumer, and therefore for society, because—assuming that economies of scale hold, then bigger is the only way—no, the One True Way—to produce more goods at a decreasing per unit cost. As we have seen, and as I will argue in future posts to this blog, lean enterprise reverses this logic. In a lean enterprise, Small is Beautiful for the consumer, and therefore for society in general. In other words, based upon the economic logic of lean enterprise, there is good reason to tighten antitrust law to prevent firms from becoming too big to fail—even Toyota, which embarrassingly enough, has joined the line for government handouts from Japan and Europe.

Tom Jackson
Portland, Oregon

Et tu, Toyota?

So, mighty Toyota is asking for handouts, a $2 billion handout from the Japanese government and a loan from the European Investment Bank. See Ian Rowley’s recent Business Week article, Auto Bailout: Et tu, Toyota?).
This is all very disappointing, especially to management consultants like myself who have championed Toyota’s approach to managing business. I suppose that we believed that, faced with an economic downturn, Toyota would simply spend a little of its famous cash reserves. But this is not any economic downturn. The scenario darkens ominously minute by minute.
Toyota’s owners are only human, more human, perhaps, than we imagined, but not inhuman. While American banks may spend their handout on dubious retention bonuses (where are all these failed bankers going to go?), Toyota will spend its handout on researching clean technologies and avoiding layoffs.
Should consultants stop preaching the Toyota Way? No. Our current economic kerfuffle has nothing to do with the Toyota Production System. The kerfuffle is a failure of markets, a failure of intellectual understanding, and a failure of political will to fix those markets. Despite what may happen to the demand for your products and services or the availability of capital, organizations that adopt Toyota’s business systems will fare far better than businesses that do not.
Most businesses are still stuffed full of Toyota’s 7 deadly wastes: overproduction, transportation, waiting, inventory, motion, overprocessing, and of course defects. And most businesses still systematically ignore the knowhow of middle and frontline managers and hourly workers, which many consultants classify as the 8th deadly waste: the waste of human creativity.
That creativity is the wellspring of the world’s eventual recovery. If only we can get the bankers to believe in it again.
Bottom line: Don’t throw Toyota’s amazing production or management systems out with the economic bathwater, however murky it might become.

Tom Jackson
Portland, Oregon

The modern corporation is dead

General Motors reinvented the modern corporation in the 1920s. Today, GM is facing the end of business as usual, to put it mildly.

Many people think that Toyota’s famous just-in-time production system is to blame. And it is a fact that GM, Ford, and Chrysler never implemented just-in-time completely, except in a few of their factories. I should know. I was a member of Ford’s Lean Advisory Board, together with a bunch of Shingo Prize winners that included the venerable John Shook. We watched in horror as Ford, under CEO Jacques Nasser, swerved sharply from its commitment to its new Ford Production System (shaped in the image of Toyota) and turned to the false God of Six Sigma.

Frankly, even if the American automakers had implemented the Toyota Production System—or, for that matter, Six Sigma—in every one of their plants, they still would not have been able to beat Toyota, with or without all their extra healthcare costs. The reason is Toyota’s secret weapon: it’s management control system.

The history of management control is fascinating, and necessary to understand before one can make any sense at all of Detroit’s meltdown. The first such system we normally learn about is the Division of Labor, in which individual entrepreneurs controlled production by farming it out in pieces to relatively unskilled workers, who displaced more highly skilled and expensive craftsmen. The Division of Labor underpinned the so-called Factory System, which began in Great Britain in the 1700s and became a linchpin of America’s meteoric economic growth.

The Factory System persisted until the mid-1800s, when the challenges of building of the transcontinental railroads necessitated a new control system, one based upon the military command-and-control model of the Prussian General Staff. In 1844, the longest railroad line operated by individual entrepreneurs in the United States, was a mere 50 miles long. After the adoption of the Prussian line-and-staff model, railroads grew to 5,000 miles long. What made the difference was not a leap in railroad technology, but a leap in management control. Big Business could operate at a great distance because of the telegraph, of course, but mainly because of standardization enforced by a Chief Executive Officer and a staff of specialists that we still rely on today: marketing and sales, finance, production, engineering…. The CEO and his staff had displaced the humble entrepreneur, and the modern corporation was born.

General Motors’ brilliant CEO, Alfred Sloan, displaced the Prussian model between 1921 and 1927 with a model of decentralized corporate divisions (Chevrolet, Buick, Pontiac, Oldsmobile, Fisher Body, Cadillac) each under the control of its own, Prussian-style leader and supporting staff. How did Sloan control his new generals? Around the CEO, Sloan installed a large staff of financial experts, who managed the divisional presidents through a system of financial targets and internal audits that evolved into today’s management accounting.

Sloan’s innovations were surprisingly powerful. Consider poor Henry Ford, who was the very image of a Big Business CEO. Ford was so busy personally controlling his famous production system that he did not have time to notice that his increasingly affluent customers were asking for more speed, more luxury, more choice—to which Ford is said to have answered, “Give them any color they want, so long as it’s black.” Sloan’s slogan was “a car for every purse and purpose.” Sloan, with his new, decentralized system and tight financial controls, had more knowledge of his customers, and more operational flexibility. Ford Motor Company nearly went out of business in 1927, and didn’t recover until after Henry Ford died over twenty years later.

Which brings us back to Toyota. Toyota’s just-in-time production system is extremely decentralized, much more so than GM’s divisions. Every worker on Toyota’s lines has the responsibility—and the power—to prevent defects from getting to the customer, even if it means stopping an entire factory! Some people say that Toyota is radically decentralized. Most people in business prefer the term “empowerment” and stay away from “radical” or “decentralized.” Just the same, Toyota’s system seems to have many more decision-makers in it than GM or Ford. That accounts for Toyota’s flexibility. But how do they keep all those newly empowered decision-makers under control?

In 1963, the Toyota Motor Company won a Deming Prize for its implementation of Total Quality Management. As a result, the company’s employees are scrupulous in their observance of detailed work standards, which results in superb operational control. As part of its TQM implementation, Toyota also implemented hoshin kanri, or policy deployment—the main subject of the website, www.hoshinkanri.biz, which hosts this blog. Inspired in part by Peter Drucker’s Management by Objectives, hoshin kanri was originally intended to ensure that quality was treated as a matter of strategic importance, which is still a very good idea. But Toyota used it in a new way: to manage its finances as well as its quality, thus inventing a new control system that integrated quality and financial management—at every level of the organization. Indeed, according to Prof. H. Thomas Johnson, Toyota doesn’t permit accountants on the shop floor. In other words, Toyota’s control system has made Sloan’s management accounting obsolete!

The modern corporation is dead, and has been for some time. Its burial is a lamentably slow process, one that will take decades—Sloan’s innovations of the 1920s did not become standard business school fare until the 1950s! It is a comprehensive process, one that affects not just the automotive industry, but every manufacturing and service industry—including, notably, healthcare. And it’s a complicated process, one that involves players in business, government, influential not-for-profits like Jim Womack’s Lean Enterprise Institute, and schools of business. Finally, it’s a fascinating process that has kept me entertained for years. I plan to share the fascination, the horror, and mirth, with my readers in my future posts.

Tom Jackson
Portland, Oregon