last bastion of German and American domination, the Lexus, Infiniti, and Acura have made incredible strides in gaining equal standing with the likes of Mercedes, BMW, Cadillac, and Lincoln. In other industries, such as home entertainment, the Japanese virtually own the market. By focusing on TQM, Japanese companies compete even when their products cost more than the competition. That's quite a change for an entire country to pull off in just 30 years.

Some people might consider TQM to be passe; nonetheless, most people have only the vaguest idea what it's about. Fortunately, although TQM is no longer talked about in the religious fashion of the '80s, Business Process Reengineering (BPR) is a direct extension of TQM. BPR is at least as much of a fad today as TQM was several years ago. So in a way, TQM lives on. This is fortunate because the fundamental principles underlying both TQM and BPR are critical to restructuring American and European businesses -both in the factory and in the office -- to make them competitive in the future.

What are TQM and BPR about? Processes. In a nutshell, organizations and their management teams are working through a fundamental conceptual shift -- a paradigm shift -- that involves thinking about work in terms of processes instead of tasks. To see why this concept is so fundamental, I'm going to look at the roots of TQM and the process improvement model behind it.

PHYSICAL FACTORIES, INFORMATION FACTORIES

Suppose that you have a factory and you need to make it better. What does better mean? More products built in less time? Higher-quality products with fewer defects? Lower manufacturing costs? Less parts inventory? More product produced per unit time and more product produced per worker? Happier workers? Happier customers? Happier managers ? All of these are aspects of a better factory.

I'm talking about factories that produce physical goods, like cars, televisions, and airplanes, of course. But can better apply to information factories, as well? An information factory is any group of workers that processes information. A front office environment that processes orders, schedules shipments, responds to customer complaints, issues bills, accepts payments, orders inventory , and keeps all the books straight is an information factory .The products it produces are shipments, bills, customer records, account files, and so on. Sales and marketing organizations can be thought of as factories, too. Those departments produce sales calls, promotions, advertisements, brochures, orders, and (hopefully) satisfied customers.

Computer organizations like to think of themselves as information factories. A couple of years ago, a brief spate of articles and papers documented experiments that originated in Japan. These experiments modeled software development directly after a factory , including reusable parts and an inventory .In addition, the experimenters converted the concept of code development into one of parts assembly. That particular concept proved to be less than totally practical; however, the idea of viewing the MlS organization as an information factory that produces systems is still very appealing to computer managers who are eager to find ways to develop better, faster, and less expensive software.

33

It is so attractive to think of these office organizations as factories because of the tremendous changes that have occurred in factory organizations -- particularly in Japan -- over the past 30 years. If cars, televisions, and airplanes can be produced so much better, why can't the rest of the organization be made to improve just as much? TQM results in better car factories, but could it result in better information factories, too? Great idea, but what exactly is TQM?

To understand TQM, go back to the list of things that make a factory better, such as more product, fewer defects, and less production time. How does that happen? The answer is to think of the entire factory as a system. A system is a collection of inter- connected parts, all tied together by a describable process. The process is the key. TQM teaches that the key to better production is improvement of the process, and the key to improvement of the process is understanding and measurement.

ASSEMBLY LINES AND BUREAUCRACIES

Until recently, the processes behind most manufacturing plants were poorly understood at best. Management over the last century has focused on tasks, often to the exclusion of processes. The bureaucracies so common to most large organizations are a direct consequence of the task orientation that originally drove the growth of the modem corporation. Task orientation is a natural outgrowth of mass production and assembly lines. Assembly lines revolve around the idea of dividing complex processes into simple steps that can be carried out repetitively by workers with little training. This assembly line model of process simplification, applied in both the manufacturing plant and in the office, led directly to today's big, slow, top-heavy organizations. understanding the route from mass production to top-heavy bureaucracy is critical to understanding the process-oriented alternative; this section looks at that route in more detail.

Assembly lines are built around specialization of a function. Take a product that has to be manufactured, divide the work over and over into smaller and smaller tasks, and when the tasks are small enough, you have an assembly line. At the turn of the century this concept -- applied to automobile production -- facilitated building cars by the millions. Unfortunately, after a large assembly line is put into place, there is no straightforward way for it to become more and more effective over time.

34

Although these two assumptions may seem somewhat removed from TOM, the connection is surprisingly intimate.

In a small organization, all three of these problems are handled straightforwardly be cause communication lines are short. information is available everywhere immediately and workers can refer decisions to appropriate parts of the company quickly and easily. In a big organization, however, these three issues are fundamental deterrents to growth if not addressed properly. The result is the uncontrollable bureaucracy.

Suppose that a customer places a large order. Usually. either a salesperson or an order-entry clerk accepts that order. After the order is written up, it's sent to the credit department to ensure that the customer has sufficient credit available. At this point two things happen. First, a credit-authorization clerk interprets company rules in a consistent way by ensuring that customers who become overextended may no longer place orders. Although simple to describe, the typical rules controlling when customers are considered delinquent. when exceptions can be made, what special approvals may be required. and so on. are generally quite complex. Having a functional specialist ensures that these rules are interpreted both consistently and correctly. Second, the credit department can ensure that the company does not become overextended. Depending on a variety of external factors. orders that might normally be processed immediately may be delayed to preserve scarce cash resources.

Next, the order moves to the scheduling department, which checks inventory, plans production, and decides when and in what sequence orders are filled. Again, functional specialists interpret policy consistently (which customers come first. when special production runs can be planned, and so on) and coordinate scarce resources (such as inventory and production capacity). When the scheduling department determines the order fulfillment date, the shipping department schedules delivery, again enforcing policy and coordinating scarce resources. The process continues through the generation of bills. recording of payment. and so on. In many companies. dozens of functional specialists may handle a typical order by the time the order is fulfilled.

35

To gain some perspective, picture the typical organization as a large pyramid with the executive team occupying the small peak and the front line employees living at the broad base. All the steps in the process of handling the order occur in the middle of this pyramid; this is the so-called middle management layer. Now, however, the function of that middle management layer is very well defined -- middle management is the place where the company stores its business rules. Of the three functions requiring multiple layers of management -- successive refinement of plans, coordination of scarce resources, and consistent interpretation of policy -- all but the first can be defined as implementation of business rules. The idea that consistent interpretation of policy is an implementation of business rules is clear. Coordination of scarce resources -- scheduling manufacturing, allocating credit dollars, managing limited inventory -- is an important expression of the organization's business rules, also. These rules determine when extra production shifts can be scheduled, how long customers should wait for limited products, how much inventory should be on hand, when the company may use third-party shippers to make up for internal delivery limitations, and so on. The way that a large organization arranges for all these business rules to be implemented in a way that can be trusted is by creating a large and sophisticated middle management layer.

Middle management plays two key roles, each of which corresponds to the last two functions: coordination of scarce resources and consistent interpretation of policy. Both functions relate directly to the amount of trust that can be placed in front-line workers. First, middle management interprets policy consistently. An important rule of accounting with a great deal of history behind it says that every important transaction should require two persons for completion. This rule neatly expresses a central dilemma of big companies: how to ensure that rules are really enforced. If salespeople can make credit decisions, won't they be too generous to their customers? If customer-service personnel are allowed to decide about product returns and emergency shipments, won't it be too easy for them to appease the customer too frequently (that is, to the detriment of cost-efficiency)? The obvious answer, suggested by both accountants and common sense, is to split the transaction. The salesperson takes the order, but the credit department decides whether to honor it. The customer service clerk receives the complaint, but only the service manager can authorize a return. In this way, complex rules can be enforced correctly, rule changes can be accommodated relatively quickly, and most important of all, front-line staff never face a conflict of interest.

In the coordination of scarce resources, the case for centralized functional specialists is even more clear. How can manufacturing be scheduled effectively except by a person who can juggle all the outstanding orders at one time? How can inventory of scarce parts be managed unless one person controls all movement in and out of storage? How does the company manage scarce credit dollars effectively unless that company processes all requests at a single location?

All of this effectively explains how the management structure evolved in most large organizations throughout the world. Without sophisticated computer systems -- and a new way of thinking about organizational effectiveness -- this style of hierarchical structure is the only way to run a large organization.

36

MAKING PROCESSES BETTER

Somewhere in the 1980s, around the time that foreign cars were redefining the automobile marketplace, American companies awoke to the issue of global competition. As one industry after another found itself under attack from within and without, management started asking questions. From those questions came the idea of taking the benchmarking concept and applying it to business. In the engineering world, a benchmark is a comparison test, run at a standardized workbench, to establish meaningful numeric comparison between two competitive designs. Applying the benchmarking concept to a business allows processes inside two companies to be directly compared, on the basis of some numeric measure, to gauge one company's performance against another. The results of benchmarking can be startling, to say the least.

In some instances, foreign competitors could build products with as little as one quarter the cost accepted as normal elsewhere. In one case reported by Michael Hammer in the Harvard Business Review, Ford Motor Company discovered a competitor running its accounts payable department with less than 10 percent of the staff required to process the same volume of work at Ford. At the same time, as customers were making clear, the quality of products coming out of the competitor's inexpensive but efficient process was substantially higher than anything seen in the past.

By the late '80s the cars, televisions, and other appliances produced in new business environments became synonymous with quality itself. How did that happen, and how could quality -- usually associated with increased cost -- actually lead to lower cost and increased efficiency?

THE QUALITY GOES IN (JUST BEFORE THE NAME GOES ON

Historically, quality has been a property determined after the fact. Products get built, on an assembly line, by workers who have virtually no control over the process they are part of. At the end of the assembly line, products are inspected for quality, and, as the Zenith jingle says, the quality goes in before the name goes on. Unfortunately, by the time the inspection takes place, it's far too late to have an impact on quality.

Modem products have many parts; at its best, inspection can test such products only in a superficial fashion. Turning a television on, taking a car for a short drive, and running a computer through a quick diagnostic test really stress very little of the overall system. The classical answer to this question is statistical: take a small number of the final products and take them completely apart to gain a picture of how deep the quality really goes. However, this overall approach to building complex products is fundamentally flawed when it comes to quality.

Suppose that in building a particular part, one of many making up a larger product, the error rate is about 1 per 100,000 or 1: 100,000. In other words, for every 100,000 parts built, only one is fundamentally flawed. Sounds like a minuscule error rate,

37

right? Wrong. In a product as complex as a car, even an error rate as low as 1 per 100,000 per part will result in every car built having at least one part with an error. How can this be?

Suppose that you roll a pair of dice, and any time you roll a five, that represents an error. With one die, errors occur one-sixth of the time; a result with no error occurs 83 percent of the time. With two dice, errors occur in 11 out of 36 rolls; the correct result comes up 69 percent of the time. With 10 dice, a roll with no errors (that is, no fives) occurs only 16 percent of the time. By the time you get to 100 dice, rolls with no errors almost never occur (.00000012 percent of the time). Of course, an error rate of 1:100,000 is much smaller than 1:6, but cars, televisions, and computers have far more than ten or a hundred parts in them. If even an error rate as low as 1:100,000 is too high, how on earth can a company build products in the real world? That's the question the TOM movement forces businesses to look at.

Consider how one bad part was built into the finished product in the first place. This happened because the product process is an assembly line. The person building that part has no intrinsic way of knowing that the part is bad because he or she is producing that part in isolation from the bigger job being carried out by the assembly line as a whole. To see how this works, and how it might work differently, imagine a mechanical device, perhaps a farm tractor with 1,000 parts, and two different ways of manufacturing such a vehicle.

The assembly-line technique calls for each of the 1,000 parts to be built individually; at the end, all 1,000 parts are assembled and the tractor is inspected on the way out of the factory. Statistically, even if each part is out of whack one time out of 100,000, this factory will almost never produce a completely correct tractor. After all, with 1,000 parts, one of them is bound to have a problem.

Now imagine a different factory that builds tractors in a very modular fashion. Each tractor has ten major assemblies, each with ten subassemblies, each in turn with ten parts. Altogether, this second factory still produces tractors with 1,000 parts; the difference is the steps they take to get there. Each subassembly is produced by an individual; as part of the production process, the individual calibrates his or her subassembly carefully, weeding out defective parts in the process and producing sub-assemblies that perform perfectly each time. In turn, each assembly is assembled by a trained individual who calibrates that assembly. Occasionally, an assembly is rejected and the subassemblies are sent back to the appropriate part of the factory; but when an assembly is done, it works and works well. Finally, the complete tractor is put together by someone working with only ten assemblies, each already properly calibrated. The chances that the worker won't be able to deal with the variations

38

introduced by only ten assemblies are pretty minimal, so the final assembly step almost always goes completely smoothly. Best of all, because each tractor is finished and calibrated by someone who can ensure that it works just right, there is no inspection process on the way out.

In the first factory, quality goes in right at the end; tractors are inspected after they are built. Indeed, quality can't go in any earlier because until the end, each tractor is just a box of parts. In the second factory, quality goes in at every step. Each subassembly, each assembly, and finally each tractor is built and calibrated to ensure perfect operation. That the entire process in the second factory is built to ensure quality is an intrinsic part of the entire production line.

PROCESS UNDERSTANDING, PROCESS IMPROVEMENT

The first rule says that assembly-line workers are robots, and the second rule says that you can tell how well the robots have done only by measuring their work and analyzing the collected information after the fact. TOM says that both these rules are wrong.

Why are robots such a problem? Well, remember that if even one part in 100,000 is wrong, almost every finished product going out the door will have some problem somewhere. The problem is that robots can't be asked to build parts that never have problems. Robots can build parts that rarely have problems, but that's not good enough. We need parts that never have problems, and robots can't do that.

Robots? Come on, isn't that an exaggeration? In Modern Times, Charlie Chaplin worked on an assembly line so repetitively that by the end of the day, his arms repeated the same motion even though the work shift had ended. That's a robot -- the same type of robot that can put a Coke can in a drive shaft and create the legends that buying a car built on a Monday or Friday is a recipe for disaster. Assembly-line

39

work may draw on the body, but it puts little emphasis on the associated brain. Factories can be exciting places to work, and they can also be spiritless machines in which people become cogs. The people are not the problem; the system is. And what TOM is about is altering that system so that factories build on the abilities of the people in them.

SELF-REGULATING PROCESSES: TURNING THE WORLD INSIDE OUT

TQM takes the assembly-line system's two rules and turns them inside out. Rather than measure quality after products are built, TQM says to just make them perfect to begin with. Rather than accept the idea that sbme parts will be wrong, establish processes that ensure that only perfect parts are ever built. How can that be? Instead of having unthinking robots, disconnected from any information about the parts they're building, have the producers of the parts become responsible for ensuring that those parts are perfect before they ever get built into the product.

The assembly-line philosophy says that the manufacturing process is very complex; simplify it by slicing it up into simple tasks. Because each task is simple, there is no way for the output of that task to be perfect. Simplicity means exactly slicing and dicing the overall manufacturing process up so fine that the tasks involve no thinking. Without thinking, there can be no quality.

What does quality mean? Quality means consistency of results. Cars that always work. Televisions that always have exactly the right colors. Not only does the car work, but also every part of the car works exactly as it's supposed to. Where does consistency come from?

In theory, it should be possible to design machines and processes that always produce the same results. In practice, there is too much variation in the real world for this to be possible. There is, however, an alternative: machines and processes that regulate themselves.

A home heating system is a good example of a self-regulating system. In the face of all the changes in the outside world, getting a furnace to maintain a completely stable temperature is a hopeless task. But with the help of a thermostat telling the furnace when to turn on and off, stable temperatures are the norm. After the thermostat is set, the heating system regulates itself.

40

More than just tuning existing production techniques, the result of TQM is to turn the conventional factory inside out. Instead of focusing on the manufacturing process just long enough to slice it into tasks, TQM demands that companies combine those tasks back into processes. Moreover, the subprocesses have to be self-regulating, so all of a sudden the picture of the typical production worker needs to be reexamined, too. Information becomes a valuable part of the manufacturing process itself rather than a by-product that can be used to determine how well the process worked after the fact. Information derived as part of the manufacturing process is now fed right back into the process -- at the time it is derived -- to keep the process self-regulating.

SELF-MANAGED PROCESSES, SELF-MANAGED TEAMS

The biggest change caused by TQM is the change to the role of the classical production worker. As the focus of the factory changes from small, self-contained tasks to larger, self-regulating processes, the role of the worker changes from unthinking robot to constantly thinking member of a self-regulating team. What is a self-regulating team?

During that period of time in the '80s when it briefly appeared that American car manufacturers might even be driven out of business by foreign imports, General Motors set about to build a completely new division. The Saturn division was charged with the mission of building a new kind of car company based on the best principles that could be discovered world-wide for building better cars. After careful study, one of the central principles the division identified was the need to eliminate the distinction between managers and workers.

In a classical factory, workers build products while managers measure how they're doing. When improvements are required, the managers divide up the unthinking tasks in new ways, tell the workers their new assignments, and hope for the best -- all after the fact. As I've discussed, the problem is that this schizophrenic process, with management and workers divided, just doesn't produce products that are good enough -- even at its best. "Nearly" perfect parts make a product that is far from perfect. Perfect parts -- always perfect, not just mostly perfect -- lead to perfect products, but robot workers and disconnected managers never produce always perfect parts.

In the process of creating itself, the Saturn division followed the example set by the Japanese and eliminated the distinction between managers and workers. Why not build a complete factory where every process is self-regulating? Why not build a factory, even in the United States, where every part is built to be perfect? A factory that produces cars that can compete with Japanese and German cars on their own terms? Are such factories possible?

The '80s, the same period during which the need for a Saturn division became evident, saw a real revolution in thinking about quality. That revolution eventually became most evident in the eyes of customers. Historically, even as recently as the '60s,

41

industrial products in general were viewed as throwaway. A television set, for example, could be expected to work for several years, start needing repairs, and eventually require complete replacement. Automobiles, toasters, lawn mowers, and other industrial consumer products were simply assumed to be intrinsically imperfect with relatively short lifetimes. Repairs and downtime were an accepted fact of life.

The profound change in both actual product quality and the perception of that quality by customers revolves completely around these self-regulating processes. Self-regulating processes require a completely new style of organization. And getting there requires a complete reexamination of all the processes inside a company -- not just the physical manufacturing processes.

The first question to ask about self-regulating processes is who is the self that is doing the regulating? The answer, obvious only after the fact, is the workers or team responsible for implementing that process. Robots, at least unthinking ones, cannot be self-regulating. Self-regulation means that the process implementors are able to modify the process while it is running. Modification of the process requires thinking.

42

Compared to the philosophy underlying assembly lines, these two rules place a great deal of responsibility on the shoulders of production and front-line workers of all kinds. Putting these rules into place requires two fundamental changes in organization -- one mechanical and the other philosophical.

The mechanical change involves rethinking the structure of the organization itself. Assembly lines, with their focus on narrowly defined tasks, make self-regulating processes virtually impossible; the specific tasks are too limited in scope to allow production workers to have enough control to guarantee perfect products. Consequently, the first step required to implement the new TQM-driven ground rules is the redesign of the fundamental production process and all the jobs in it. The scope of that redesign is all-encompassing and reaches farther than this book can deal with in detail. If you're in the mood to stimulate your thinking about this topic, however, a good book is The Goal, by Eleyahu Goldratt and Jeff Cox.

One example illustrates the fundamental nature of the shift. In an assembly-line world, manufacturers and their suppliers are locked into an adversarial relationship. The manufacturer buys parts at the lowest possible price, and the suppliers compete with each other for the manufacturer's business. The market is a jungle, and the fittest survive. Do whatever is required to win the business; do whatever is required to get the lowest possible price from your suppliers. That's the world most managers, marketers, and salespeople know. If the entire production process is to be self-regulating and if every part must be perfect, then clearly external suppliers are an intrinsic part of that picture. If suppliers are to be self-regulating, they need access to a great deal of information about product requirements, performance of the parts they supply, and so on. In fact, those suppliers are part of the team that incorporates their parts into the actual product. If they're part of the team, how can there be an adversarial relationship? And if they're not part of the team, how can the process they're part of be self-regulating? There it is: the revolution in progress. Suddenly suppliers, historically almost enemies, are not just friends but integral parts of the bigger family that builds the product. They're not just not enemies, not just friends -- they're members of the family.

This example illustrates how far the organizational redesign required by TQM really extends. It also illustrates another fundamental aspect of that redesign: a focus on larger processes instead of smaller tasks. A basic consequence of that shift is a focus on teams instead of individual workers. Many times the task performed by an individual production worker can be enlarged enough so that that worker can be

43

responsible for a self-regulating process. Instead of tightening a single hinge all day, a worker assembles complete doors and tests them to ensure perfection. Other times, in order to enlarge that task to become a self-regulating process, a team has to be assembled. Instead of installing just a spark plug, the worker becomes part of a team that assembles and tests complete engines.

All of this leads to the second fundamental organizational shift associated with TQM: a philosophical shift more profound than the mechanical shift I just described. The philosophical shift involves converting workers who historically have been unthinking assembly-line robots into thinking owners of self-regulating processes. Redesigning jobs and organizations is an intellectual process. Perhaps that process is demanding and intricate, but given smart enough people with enough experience, that intellectual process is quite straightforward. Converting unthinking robots into owners of self- regulating processes, on the other hand, is a cultural shift. That cultural shift is now the primary challenge confronting TQM, BPR, and as it turns out, client/server.

What was the biggest challenge facing the Saturn project? Not redesigning the factory, not developing engineering plans for a new type of car, and not deciding where to build the physical plant. The biggest challenge was empowering the employees so that they could participate in self-managed teams. If every engine, every transmission, every dashboard had to be perfect every time, the teams producing those engines, transmissions, and dashboards needed to have a new attitude. The mindset that I'm here from 9 to 5, and management can worry about it if defect rates get too high or the process isn't quite right had been proven not to work. Instead, every worker on every team had to believe that he or she was responsible for ensuring that every item built is perfect and that nobody else will fix problems if he or she doesn't. This set of values and beliefs is very different from what is normal in most manufacturing plants. What's being described is no less than a cultural revolution.

Instituting this kind of cultural turn requires a great deal of work, significant investment, and plenty of time. At Saturn and at other companies around the world, workers, managers, and supervisors found themselves participating in Outward Bound courses -- walking tightropes and jumping from high poles. More than just trying new things together, supervisors, managers, and workers had to start learning to know each other as people, understanding each other's needs, and supporting each other as part of the same team. Tough stuff, but the alternative, not responding to global competition, is tougher.

THE OFFICE AS FACTORY -- BUSINESS PROCESSES VERSUS BUSINESS TASKS

If an organization is to consist entirely of self-regulating processes, those processes have to reach far beyond the factory floor. Here are some important reasons why:

44

Products can't be built on time with inventory that doesn't exist. Incorrect specifications can cause just as many wrong parts as teams that don't build parts right. So it's not surprising that, after TOM had started revolutionizing the manufacturing environment itself, the next step was for the same philosophy to start finding its way into the rest of the organization.

TOM encourages us to think about broad processes instead of individual tasks. When the same thinking is applied to information and office processes, the results are surprising. Quality is often thought of in terms of defects and product performance. Productivity, however, should not be overlooked. Productivity means not only processing orders correctly, but also focusing on the amount of time it takes to process orders.

In fact, the classical complaint about information-based procedures generally deals with time, not accuracy. My order took three weeks! The company took nine weeks to handle my complaint! How can a loan application take 17 days? Why are bureaucracies so slow and inflexible? That's what everybody wants to know. The answer revolves around tasks.

Office procedures, much like manufacturing processes, historically have been designed around the philosophy of the assembly line. Complex processes are reduced to very small tasks that can be carried out by unskilled, unthinking workers. All important decisions are funneled to the middle of the organization where the business rules are stored. This organizational structure has two basic consequences.

First, step-by-step, task-oriented information processes are incredibly slow. Each task is always associated with a queue. As an order, a complaint, or an application travels through the organization, it spends more time in queues than it does being processed.

Task-oriented information processes are also inflexible. The biggest complaint about bureaucracies is that they are both slow and inflexible. How often have you been frustrated because a so-called customer service representative says he or she can't change the rules? On the flip side, how do service-oriented organizations create a reputation for outstanding service? By empowering their front-line employees to make decisions. Nordstrom, the clothing retailer, has built a world-wide reputation for service and the customer loyalty that goes with it by empowering its sales staff to make the decisions required to meet the customer's needs. In one mythical story, a customer service representative allowed a customer to retum a pair of snow tires, even

45

though Nordstrom sells nothing to do with cars. The value in terms of reputation gained from employees and teams empowered to ensure customer satisfaction far outweighs the costs associated with potentially suboptimal decisions.

Another example, cited by Michael Hammer and James Champy in Reengineering the Corporation (Harper Business Books, 1993), illustrates the change in philosophy and practice behind reengineering business processes. Recall the three problems as sociated with size; the problems solved by the classical bureaucracy: planning, coordination, and policy enforcement. IBM Credit Corporation historically processed credit applications in a little over seven days. Each application was handled by five different specialists, each in a different department, to ensure that policies were enforced consistently and that the scarce credit resource was coordinated effectively. By establishing a single team to handle each request and empowering it with the authority and responsibility for rule enforcement, loan application time was trimmed from seven days to four hours. Trimmed? Seven days to four hours? By moving responsibility for business rules out to the front line, the credit approval process was more than trimmed; it was radically redefined.

Business Process Reengineering involves applying the same critical eye that works in the factory to the procedures and processes found in the office environment. As in the factory, the shift is fundamental and simultaneously mechanical and philosophical. The mechanical shift involves rethinking all the processes found in an organization to make them all self-regulating. Along the way, business rule enforcement and coordination decisions get moved out to the front line. The philosophical shift involves convincing office workers (who were previously unthinking and had no responsibility for their results) to start thinking and taking total responsibility for their results.

Take a look at one last example of the impact that reengineering can have and then look at its impact on the classical bureaucratic model. Then, Chapter 3 lays out the technical requirements of TQM, BPR, and process orientation.

THE VIRTUAL SALES OFFICE

46

Contradictory objectives? Yes, but the paradigm shifts so often beloved by planners and consultants alike are often derived directly from such contradictory requirements. After more investigation, the management team developed a picture of sales costs that revealed that one of the largest components of the cost picture was the physical sales offices -- the bricks and mortar with desks inside. Studying the function of the sales offices and customers' feelings about the sales force, management saw a surprising picture.

Beyond their social function, the sales offices in this company had the primary function of data communication. Salespeople visited sales offices to enter orders at terminals and to print territory reports about their customers at printers located at the sales offices. How did customers feel about all those orders entered at the terminals?

Mad as heck. When asked what they wanted most in the way of improved customer service, most customers responded with a simple request: process my orders faster. A prototypical story tells the rest. A small customer ordered some products from his salesman specifying next-day delivery; the customer needed the product for an installation scheduled for the day after that. The salesman promised to meet the request and drove specially to his office to enter the order. Like always, the order was processed through four different departments. In spite of being marked "urgent" and requiring next-day delivery, the order made it through only two departments the first day. On arriving at the third department, the shipping scheduler interpreted company policy narrowly and decided the order wasn't big enough to warrant expedited delivery. The fourth department, responsible for (among other things) customer satisfaction, realized that the terms of the order had been altered, but because of the small size of the order, elected to notify the customer by regular mail. The result? Three days after placing the order, the customer was notified by mail that his order would arrive the following week as part of a regular shipment. One set of rules interpreted consistently led to one less customer to worry about in the future.

Today the company has shut down almost all of its physical sales offices. Each sales-person has been provided with a notebook computer equipped with a cellular modem. The savings associated with shutting down the sales offices allowed the company to grow the sales force by 5 percent while still reducing net sales cost by 3 percent. Best of all, each notebook computer contains an order processing application that under- stands all the company's key business rules. That notebook communicates with regional servers that manage inventory , schedule shipments, and can commit to guaranteed delivery dates.

Each salesperson has both a sales target and a profitability target for his or her territory and customer base. Each decision to expedite a shipment, each decision to honor an unusual return request, and each potential exception to the default rules can

47

be evaluated on the spot in terms of its impact on profitability. Of course, the salesperson can make his or her own subjective judgment about the impact on future sales. Best of all, though, when the salesperson promises next-day delivery, the promise has value and meaning.

The new result? The customers are amazed. The products they buy are still solid in terms of price and features, but for the need they fill, that middle-of-the-road positioning is fine. More important, the big, old company suddenly offers better service, and best of all, promises customers can count on. Not only that, but as the same customers say in unsolicited letters, they get previously unimaginable flexibility. As those letters say, the salespeople truly represent their company now; they don't just receive orders that are processed elsewhere; instead, they are the company for their customers.

BUREAUCRACY: A NEW MODEL

Applied to both the factory and the office, TQM and BPR represent a new model for thinking about companies and the people who work for them. Much more than just a clever technique for improving efficiency, these two approaches call for a fundamental cultural shift that promises to enrich jobs, produce better products at lower costs, and finally make bureaucracies flexible after all. So why do we even need bureaucracies?

Recall again the three fundamental roles of the bureaucracy: planning, coordination, and policy enforcement. Based on these three roles, what is the smallest bureaucracy really needed? There is an answer.

Elliott Jaques has studied organizations for over 40 years. Writing in the Harvard Business Review in 1990, he asserts that the fundamental defining characteristic of organizational size, other than number of employees, is hierarchy. In a hierarchical organization, there is a leader at the top of the hierarchy. In each successive layer of the hierarchy, managers have other employees reporting to them. First invented by the military , hierarchical command and control is a fine structure for organizing large numbers of people around a common purpose.

Jaques reminds us of the importance of the third function of a hierarchy -- planning -- which is the first role of a bureaucracy. That function deals with the development of long-range plans and the translation of those plans into tactical action plans, one year at a time. Based on his research, Professor Jaques even tells how many layers of hierarchy are required -- at a minimum -- to deal with this planning and translation of vision into function. The model developed by Jaques deals with time frames.

How far ahead is a worker at each level of the hierarchy required to plan in doing his or her normal job? More controversially, what is a fair level of pay (in 1990 dollars) for a worker with the responsibility for looking ahead a certain distance into the future? The answers appear in Table 2-1. Smaller organizations, those with no requirement or ability to plan more than two years into the future, can make do with four levels of hierarchy. Of course, a really small company, working from day to day, doesn't need even that much. However, even the very biggest company, planning 20 years into the future, should be able to function very effectively with only seven layers of hierarchy. How can that be? Don't typical big companies have dozens of levels?

48

Yes, companies typically have many more layers of hierarchy and bureaucracy than this model suggests. These companies also have cultural models that suggest that no manager can have more than 7 to l0 direct reports -- the classical span of control. And finally, these organizations operate in climates with low levels of trust and minimal empowerment, which creates a need for complex, top-heavy bureaucracies to implement all the business rules that keep the company running.

Table 2-1.  A Model for Bureaucratic Hierarchy

Recently, advocates and advanced practitioners of BPR, TQM, employee empowerment, and self-managed teams have begun suggesting that the classical span of control may be too limited. In an organization with self-regulating processes and truly empowered employees, a manager may potentially be capable of handling larger numbers of employees reporting to him or her. The same movement also suggests that the central bureaucracy responsible for enforcing business rules and coordinating resources not only is not required, but actually is a cause of poor performance.

A new model for the future is emerging. Hierarchical organizations are not disappearing. What may be disappearing is the bureaucracy normally associated with both size and hierarchy. To put this in perspective, consider the American Heritage Dictionary's definition of the term bureaucracy:

1.   a. Administration of a government chiefly through bureaus staffed with nonelective officials. b. The departments and their officials as a group.
   
2. Government marked by diffusion of authority among numerous offices and adherence to inflexible rules of operation.
   
3. An administrative system in which the need to follow complex procedures impedes effective action.