F IGURE 2.3: T ESCO R EORDER S YSTEM
As a result of this system of daily replenishment, Tesco has increased the “service level” to its retail stores (the percentage of supplier shipments which arrive exactly on time in good condition in exactly the right amount) from 92 to 98.5 percent. At the same time, the stocks on hand of the average good (in the retail stores plus the RDCs) fell from 21 to 12.8 days. For “fast movers” like cola, accounting for more than half of Tesco’s total sales, inventories at the RDC and the retail store combined are now only 3 to 5 days.
However, as Tesco did this, they learned the limits of what can be accomplished by one firm alone. Specifically, first-tier suppliers like its bottlers have been fulfilling Tesco orders nightly, just-in-time, but from massive finished goods inventories. Their production methods—with high-speed machines, long changeover times, and large batches—have given them no real choice. 15 Meanwhile, the firms farther up the value stream from the bottler, also using massive high-speed machines with long changeovers to produce large batches, have not yet even taken the step of delivering just-in-time from finished goods inventories. Because the bottler cannot get rapid response from its upstream suppliers to changing levels of demand, it continues to order batches of goods at weekly, monthly, or even quarterly intervals (in the case of some raw materials).
If Tesco wants to shrink costs and improve the reliability of the 85 percent of the value stream it does not directly control, it’s obvious that the upstream firms must collectively rethink their operating methods, and this is how Tesco and the Lean Enterprise Research Centre joined forces. While it is still in the early stages, the process of jointly conducting the analysis just described should gradually change Tesco, the bottler, the can maker, the cold roller, the hot roller, the smelter, and the bauxite miner from seven isolated adversaries into a team of collaborators, indeed into a lean enterprise.
Creating Cola
The final element in the cola story is the value stream for product development. Historically, in the grocery business, first-tier suppliers like the bottler or the branded purveyor of goods have been responsible for the great bulk of product innovations and introductions. Yet only a brief effort to list the activities in the value stream culminating in the launch of a new product raises many questions.
Typically, a firm like the bottler is continually looking for new products to defend its current market share, to expand its scope of offerings (and justify more shelf space at Tesco), and to substitute products with higher margins for old standbys like cola. In the industry, the typical product development cycle is about one year and consists of a number of product clinics followed by larger product trials culminating in the decision for a full-scale launch.
Although the actual steps involved are very simple and typically involve very little true “research and development,” they are conducted sequentially so that if one looks down on a product concept from a bird’s-eye view it is quickly apparent that during most of the development period the concept is sitting still, awaiting feedback from the group which conducts the clinics on all of the firm’s products or awaiting its place on the schedule of the department which conducts small-scale market trials for all products. Then, when the decision to launch is made, there is more waiting while the production system is adapted to accept the new product, new packaging materials are developed, and the marketing campaign is planned.
The end result of this system is that new products—which are often “new” only in the sense of having reformulated ingredients (for example, caffeine-free and cherry cola)—cost an average of $15 million to launch (half of this going to advertising) and … usually fail in the marketplace. 16
The result for Tesco is large amounts of shelf space tied up with “new” products that don’t sell and are launched at the same time in the stores of its direct competitors. The obvious question is: How can it take a year’s development time and a $15 million expenditure to introduce a “new” product which isn’t new and which no one wants?
Simply reducing development time and expense, while highly desirable, will not be enough to have much effect on this value stream, so Tesco has started to rethink the product development process on a more fundamental level in terms of value. Perhaps, just as the individual steps in the value stream are incomprehensible in isolation, customers do not really want to shop for isolated items. Would it perhaps be better for Tesco and its bottler to jointly undertake the development of the full complement of beverages necessary to keep Tesco customers happy, and for Tesco to develop longer-term relations with its customers so they would not be strangers? Toward this end Tesco has just launched a frequent-shopper program that will gather purchase pattern data on every regular customer and should permit a more coherent value stream in product development.
Putting Value Stream Analysis to Work
Having looked at the specific steps involved in the value stream for one specific product, we are ready to put our findings to work more broadly. In the cola case, unlike the Pratt & Whitney example cited in the Introduction, we do not see any steps in the third category which can be immediately eliminated because they are simply redundant. Instead, we see a large number of steps in the second category. They clearly add no value—they’re muda —and they therefore become targets for elimination by application of lean techniques.
Note that in performing this analysis we are not “benchmarking” by comparing Tesco’s cola value stream with those of its competitors. Although we gave a boost to the benchmarking industry with our previous book, The Machine That Changed the World, which described the most comprehensive benchmarking ever attempted in a gigantic global industry, we now feel that benchmarking is a waste of time for managers that understand lean thinking. 1 7
Lean benchmarkers who discover their performance is superior to their competitors’ have a natural tendency to relax (the risk Tesco would run today in benchmarking its internal operations) while mass producers discovering that their performance is inferior often have a hard time understanding exactly why (for example, General Motors and Volkswagen in the 1980s). They tend to get distracted by easy-to-measure or impossible-to-emulate differences in factor costs, scale, or “culture,” when the really important differences lie in the harder-to-see ways value-creating activities are organized.
Our earnest advice to lean firms today is simple: To hell with your competitors; compete against perfection by identifying all activities that are muda and eliminating them. This is an absolute rather than a relative standard which can provide the essential North Star for any organization. (In its most spectacular application, it has kept the Toyota organization in the lead for forty years.) However, to put this admonition to work you must master the key techniques for eliminating muda. It all begins with flow.
CHAPTER 3
Flow
The World of Batch-and-Queue
What happens when you go to your doctor? Usually, you make an appointment some days ahead, then arrive at the appointed time and sit in a waiting room. When the doctor sees you—usually behind schedule—she or he makes a judgment about what your problem is likely to be. You are then routed to the appropriate specialist, quite possibly on another day, certainly after sitting in another waiting room. Your specialist will need to order tests using large, dedicated laboratory equipment, requiring another wait and then another visit to review the results. Then, if the nature of the problem is clear, it’s time for the appropriate treatment, perhaps involving a trip to the pharmacy (and another line), perhaps a trip back to the specialist for a complex procedure (complete with wait). If you are unlucky and require hospital treatment, you enter a whole new world of specialized functions, disconnected processes, and waiting.
If you take a moment to reflect on your experience, you discover that the amount of time actually spent on your treatment was a tiny fraction of the time you spent going through the “process.” Mostly you were sitting and waiting (“patient” is clearly the right word), or moving a
bout to the next step in the diagnosis and treatment. You put up with this because you’ve been told that all this stopping and starting and being handed off to strangers is the price of “efficiency” in receiving the highest-quality care.
We’ve already looked briefly at another service, a trip involving an airline. And most of the time the experience is even worse than the Joneses’ family trip to Crete because rather than taking a direct flight you must go through a hub for sortation. In the end, the time you spend actually moving along the most direct route is likely to be little more than half the total time required to get from door to door. Yet most travelers put up with this system without dreaming of anything better. After all, it’s extremely safe, and travelers are told that it’s highly efficient because it fully utilizes expensive airplanes and airports.
Health care and travel are usually called “personal services,” in contrast with “products” like VCRs, washing machines, Wiremold’s wire guides, and Tesco’s beverages. Actually, the major difference is that in the case of health care and travel, you the customer are being acted upon—you are necessarily part of the production process. With goods, by contrast, you wait at the end of the process, seemingly beyond harm’s reach. However, there is no escaping the consequences of the way the job gets done even if you are not directly involved.
Let’s take just one example for a common good, the single-family home. Henry Ford dreamed about mass-producing homes using standard but modularized designs with the modules built in factories to slash design and production costs while still providing variety. A number of entrepreneurs actually created modular designs and briefly set up production lines in the United States to make the modules for prefabricated houses immediately after World War II. 1 And Toyota has had modest success in Japan since the 1960s in offering a wide range of floor plans and exterior appearances using a few basic modules fabricated on a production line and assembled almost instantly at the construction site.
Yet, almost all of the world’s new single-family homes are still built largely at the construction site by cutting and fastening a welter of materials to create the basic structure and then installing thousands of individual components, from plumbing fixtures to kitchen appliances to wall sockets.
If you go to your home builder and then to the construction site and take a seat to watch the action, you will mostly note inaction. For example, when Doyle Wilson started to measure what occurred in his office and at the work site as part of his TQM effort, he discovered that five-sixths of the typical construction schedule for a custom-built home was occupied with two activities: waiting for the next set of specialists (architects, cost estimators, bill-of-material drafters, landscape architects, roofers, sheetrockers, plumbers, electricians, landscapers) to work a particular job into their complex schedules, and rework to rip out and correct the work just done that was either incorrect from a technical standpoint or failed to meet the needs and expectations of the home buyer.
As the buyer at the end of the process, you pay for all the waiting and rework—grumbling, of course—but it is a custom product, after all, and you’ve heard many stories from your friends about even worse problems with their homes, so you tend to accept the predominant system and its problems as unavoidable and inherent to the nature of the activity.
In fact, all of these activities—the creation, ordering, and provision of any good or any service—can be made to flow. And when we start thinking about ways to line up all of the essential steps needed to get a job done into a steady, continuous flow, with no wasted motions, no interruptions, no batches, and no queues, it changes everything: how we work together, the kinds of tools we devise to help with our work, the organizations we create to facilitate the flow, the kinds of careers we pursue, the nature of business firms (including nonprofit service providers) and their linkages to each other and society.
Applying flow to the full range of human activities will not be easy or automatic. For starters, it’s hard for most managers to even see the flow of value and, therefore, to grasp the value of flow. Then, once managers begin to see, many practical problems must be overcome to fully introduce and sustain flow. However, we do insist that flow principles can be applied to any activity and that the consequences are always dramatic. Indeed, the amount of human effort, time, space, tools, and inventories needed to design and provide a given service or good can typically be cut in half very quickly, and steady progress can be maintained from this point onward to cut inputs in half again within a few years.
The Techniques of Flow
So, how do you make value flow? The first step, once value is defined and the entire value stream is identified, is to focus on the actual object—the specific design, the specific order, and the product itself (a “cure,” a trip, a house, a bicycle)—and never let it out of sight from beginning to completion. The second step, which makes the first step possible, is to ignore the traditional boundaries of jobs, careers, functions (often organized into departments), and firms to form a lean enterprise removing all impediments to the continuous flow of the specific product or product family. The third step is to rethink specific work practices and tools to eliminate backflows, scrap, and stoppages of all sorts so that the design, order, and production of the specific product can proceed continuously.
In fact, these three steps must be taken together. Most managers imagine that the requirements of efficiency dictate that designs, orders, and products go “through the system” and that good management consists of avoiding variances in the performance of the complex system handling a wide variety of products. The real need is to get rid of the system and start over, on a new basis. To make this approach clear and specific, let’s take as a concrete example the design, ordering, and production of a bicycle .
From Batch to Flow in Bicycles
We’ve chosen this example partly because the bicycle itself is simple and lacks glamour. You will not be distracted by novel product designs or exotic technologies. We’ve also chosen it because we happen to know something about the bicycle industry, one of us having resolved to test the methods we describe in this book by taking an ownership position in a real bicycle company. Finally, we have chosen bicycle manufacture because it is a deeply disintegrated industry, with most final-assembler firms making only the frame while buying the components—wheels, brakes, gears, seats, handle-bars, plus raw materials in the form of frame tubing—from a long list of supplier companies, many larger than the final assemblers themselves. The problems of value stream integration are present in abundance.
DESIGN
Product design in the bicycle industry was historically a classic batch-and-queue affair in which the marketing department determined a “need,” the product engineers then designed a product to serve the need, the prototype department built a prototype to test the design, the tooling department designed tools to make a high-volume version of the approved prototype, and the production engineering group in the manufacturing department figured out how to use the tools to fabricate the frame and then assemble the component parts into a completed bike. Meanwhile, the purchasing department, once the design was finalized, arranged to buy the necessary component parts for delivery to the assembly hall.
A design for a new product, usually only one of many under development at a given time, moved from department to department, waiting in the queue in each department. Frequently it went back for rework to a previous department or was secretly reengineered at a point downstream to deal with incompatibilities between the perspectives of, say, the tool designers and the product designers who handled the design in the previous step. There was no flow.
In the late 1980s and early 1990s, most firms switched to “heavyweight” program management with a strong team leader and a few dedicated team members, but without changing the rest of the system. The product “team” was really just a committee with a staff that sent the great bulk of the actual development work back to the departments, where it still waited in queues. What’s more, th
ere was no effective methodology for carrying designs through the system without lots of rework and backflows. Even worse, no one was really responsible for the final results of development efforts because the accounting and reward systems never linked the success of a product through its production life with the original efforts of the design team. There was, therefore, a bias toward ingenious designs with admirable technical features which customers liked but which failed to return a profit due to excess costs and launch delays.
The lean approach is to create truly dedicated product teams with all the skills needed to conduct value specification, general design, detailed engineering, purchasing, tooling, and production planning in one room in a short period of time using a proved team decision-making methodology commonly called Quality Function Deployment (QFD). 2 This method permits development teams to standardize work so that a team follows the same approach every time. Because every team in a firm also follows this approach, it’s possible to accurately measure throughput time and to continually improve the design methodology itself.
With a truly dedicated team in place, rigorously using QFD to correctly specify value and then eliminate rework and backflows, the design never stops moving forward until it’s fully in production. The result, as we will demonstrate in the examples in Part II , is to reduce development time by more than half and the amount of effort needed by more than half while getting a much higher “hit rate” of products which actually speak to the needs of customers.
In our experience, dedicated product teams do not need to be nearly as large as traditional managers would predict, and the smaller they can be kept the better all around. A host of narrowly skilled specialists are not needed because most marketing, engineering, purchasing, and production professionals actually have much broader skills than they have (1) ever realized, (2) ever admitted, or (3) ever been allowed to use. When a small team is given the mandate to “just do it,” we always find that the professionals suddenly discover that each can successfully cover a much broader scope of tasks than they have ever been allowed to previously. They do the job well and they enjoy it.
Lean Thinking Page 6