Lean Thinking

by Daniel T Jones

  We are also seeing many managers who use the current era of stagnation as an excuse to abandon any efforts to improve their current operations. Instead they relocate design and production, almost always using mass production methods, to remote locations, often thousands of miles from their customers. The new locations have one key attribute—low factor costs, particularly in the form of cheap labor—and they seem irresistible.

  The problem is that every competitor can immediately pursue the same strategy, so the advantage is short-lived. In addition, because firms following this strategy add nothing to their knowledge of lean practice, they are vulnerable to shifts in currency rates and geopolitics that may require them to move again soon. We’ll return to this point in a moment in discussing the need to optimize entire value streams on a global basis.

  MAP YOUR VALUE STREAMS

  This has been the greatest area of learning for us because we had not grasped just how much help the average manager needs to see the value stream. The maps we drew on pages 39 and 42, although accurate and provocative, turned out to be too simple. And we made a critical error by failing to connect on one map the flow of information going back from the customer to the producer with the transforming actions on the product, in response to this information, as the good or service moves toward the customer. Making this connection is the critical leap in being able to see the closed circuit of demand and response that is the essence of value creation, an insight that traditional process maps, showing physical transformations alone, fail to provide.

  Thus we are deeply grateful to Mike Rother and John Shook for adapting Toyota’s standard method for portraying material and information flows into the value stream maps we now use. 2 These maps can be drawn at any scale, from a simple administrative process within an office to the global flow of an extended value stream running from raw materials in the ground to the end consumer.

  The objective in each case is to write down all of the steps in the process as it currently operates to define what we call the Current State. For each step we urge managers to ask a set of very simple questions. Does the step create value for the customer? Is the step capable? (That is, does it produce a good result every time?) Is it available? (That is, can it produce the desired output, not just the desired quality, every time?) Is it flexible? (Can it be changed over quickly from one product to the next so that items can be produced in small lots or even lots of one?) Is capacity for the step adequate so the product doesn’t need to wait on the process? Or is there too much capacity (due to designing equipment in large increments of capacity based on demand forecasts that are often wrong)? 3

  Steps that do not create value should, of course, be eliminated, while steps that are incapable, unavailable, inflexible, inadequate, and under- or overcapacitized should be perfected. But this step-by-step analysis provides only part of the picture because the relation between the steps is equally important. Does the information coming back from the customer flow smoothly without delays? Does the product moving toward the customer flow smoothly from step to step so that total throughput time is only slightly more than the sum of the times needed for individual processing steps? Does the product flow at the desire of the customer rather than at the push of the producer? Finally, is demand “leveled” at each stage so that small perturbations are smoothed rather than amplified?

  By writing down all the steps as a team, as shown in Figure 15.1 for the flow of value within the walls of a factory, it’s possible for everyone to see the whole value stream under discussion and to agree on its current level of performance.

  The map below shows the flow of information from the customer to the various points in the production process, moving from right to left in the upper half. Orders go from the customer to a Material Requirements Planning computer, where they are held in inventory awaiting the weekly run of the system to devise the production schedule for the following week. A considerable amount of information expediting occurs as floor managers discover shortages or customer demand suddenly changes.

  F IGURE 15.1: C URRENT S TATE V ALUE S TREAM M AP

  The map also shows the flow of products from raw materials to customer, moving from left to right in the lower half. It summarizes the performance of the five necessary steps, shows the inventory currently accumulating between them, compares value-creating time (very small) with total throughput time (very large), and helps managers envision the initial flow kaizen needed to drastically compress the throughput time for the product, eliminate wasted steps, and rectify quality, flexibility, availability, and adequacy problems.

  All of this information can be summarized in a box score of Current State performance as shown in Figure 15.2 .

  F IGURE 15.2: C URRENT S TATE B OX S CORE

  The visioning process facilitated by the map and the box score should lead to a vastly improved Future State, as shown in Figure 15.3 . Reaching this state requires achieving the “kaizen bursts” on the Future State map, which show the necessary points for flow and process kaizen.

  In this case, the specific steps required are to improve the capability (first-time quality), availability (uptime), and flexibility (changeover time) of the four weld and assembly steps and to eliminate the inventories impeding flow by turning the four steps into a cell. (Note that one fewer operator is required as a result.) In addition, setup times of the stamping press are greatly reduced to permit the production of much smaller batches, further reducing inventories.

  F IGURE 15.3: F UTURE S TATE V ALUE S TREAM M AP

  The final step is to disconnect the Material Requirements Planning system previously giving production orders to every step in the process. A simple pull system is put in its place that sends kanban signals from a heijunka box (a demand leveling device) at precisely paced intervals to the weld/assembly cell, which is the “pacemaker process” for this value stream. Additional pull loops are installed from the weld/assembly cell to the stamping machine and from the stamping machine to the supplier of steel coils. As a result, the entire process of information management is vastly simplified and transitioned from push to pull.

  The implications for performance of this Future State are shown in the expanded box score contrasting the Current State and the Future State in Figure 15.4 .

  The mapping process clearly reveals the potential for a major leap in performance if a relatively small number of flow and process kaizens can be conducted and then sustained. And this is not the end of the potential for improvement. As we will show below, in the section on perfecting the value stream, it is always possible to make further progress by designating the Future State, once achieved, as the new Current State and beginning the improvement cycle again.

  This brings us to our major concern about value stream mapping. We’ve found overwhelming acceptance of this tool across the world 4 and we now find many managers with beautiful Current State maps and with equally beautiful Future State maps indicating the potential for major leaps in performance. But, when we take a walk along the value stream, there is no actual Future State. The promised leap in performance has never occurred or has been achieved to only a fraction of the extent possible. 5

  F IGURE 15.4: C URRENT TO F UTURE S TATE B OX S CORE

  When we see this situation we always ask to see the plan for achieving the Future State, which should look something like Figure 15.5 , and we ask to meet the individual responsible for managing and improving the value stream. And this is the great problem: usually there is no real plan, or at least no implementable plan, because no one has the responsibility. There is no value stream manager to perfect the process.

  REORGANIZE YOUR FIRM BY PRODUCT FAMILY AND VALUE STREAM

  Just as we underestimated the importance of the value stream map, we also failed to grasp the significance of the value stream manager. This is the person who leads the mapping process and takes responsibility for removing the muda from the value stream for a product, while introducing flow and pull. Instead of describing the role of this critica
l individual in detail (whom we did mention briefly as the Directly Responsible Individual at Lantech, the Product Team Leaders at Wiremold and Pratt & Whitney, and the Chief Engineer at Toyota), we concentrated on changing the organization of the firm so that all of the needed skills within functional areas would be directly under this person’s authority.

  F IGURE 15.5: I MPLEMENTATION P LAN

  We’ve subsequently found in a number of organizations that to get the attention of self-absorbed functions, it can be helpful to change reporting arrangements and move personnel under a product line manager or team leader, at least for one product generation. We’ve also realized that Toyota and more mature lean firms (now including Lantech) get brilliant results from giving the value stream manager complete responsibility for the value stream and the success of the product but hardly any direct reports or traditional authority.

  Instead, the value stream manager develops the vision for the product, determines the Current State of the value stream, and then envisions the Future State. She or he then treats the functions as the suppliers of the essential inputs (for example, engineering, operations, purchasing, sales, lean knowledge) needed to reach this state. If the functions fail to perform, the value stream manager typically goes directly to the CEO, the COO, or the director of the office of value stream managers, to describe the problem, get to the root cause, and install a fix.

  Finally, we’ve discovered that these value stream and product line managers, like so much in the lean world, are “fractal.” That is, a product line manager overseeing an entire product may work with a number of value stream managers at lower levels taking responsibility for different courses of the value stream. For example, a chief engineer (to use Toyota’s term for a product line manager overseeing an entire automotive platform) works with a development leader in design, a value stream manager in the assembly plant, and value stream managers in each of the component plants working on major items assembled into the finished product. Each manager is essentially doing the same job but with varying scope—wide at the top and narrow at the bottom.

  That this approach works for maturing lean firms besides Toyota became apparent to us in talking with Pat Lancaster at the trade show where he launched his new right-sized wrappers (described in Chapter 14 ). We asked how Lantech’s dedicated product teams were functioning and got a quizzical look.

  “Actually, we’ve found that the Directly Responsible Individual [the value stream manager] is the critical player in our organization. Once the functional departments got enlightened, we found we no longer needed to change the organization chart and move people onto product teams for each new product family. Instead, the DRI explains to the functions what they need to do as his suppliers to ensure the success of the product. And they do it.”

  CREATE A LEAN PROMOTION FUNCTION

  In the first edition of Lean Thinking, we proposed that a lean promotion function be created to house the functional expertise from old-fashioned industrial engineering, quality, and maintenance departments along with the newfound knowledge about flow and pull. Many readers found this suggestion problematic because they could not imagine that experts from a quality background, a total productive maintenance (TPM) background, and a lean (TPS) background could work in harmony.

  As time has passed and we have listened to many pointless arguments between sensei from TPS, TPM, TQC, and TQM backgrounds, we’ve become even more certain that all of the expert animals in the lean zoo should live in the same cage. That’s because all of these experts—once differing vocabularies 6 and professional rivalries are stripped away—are in pursuit of the same goal: the perfect process.

  Each expert would like to create value streams where every step is valuable, every step is capable (the starting point of quality experts), every step is available (the point of origin of maintenance experts), every step is adequate (with neither too much nor too little capacity), and all steps are highly flexible and linked by pull and flow with leveled demand (the starting point of TPS experts). And every value stream manager would like to be able to tap a single supplier of the knowledge needed to achieve perfection.

  The challenge is to create a dialogue between all the experts so the value stream manager gets consistent, quality advice in a single voice. Only in this way can the rate of improvement can be maximized.

  The Lean Promotion Office should be small, except for periods when excess employees from line jobs are being redeployed and put to work on short-term kaizen projects. It needs only a few experts who are willing to master all of the knowledge and methods needed to create perfect value streams and to teach this knowledge, as necessary, to value stream managers and line employees. And it may get smaller over time.

  After all, lean knowledge is most needed early in the transformation when most value stream managers lack critical knowhow and the value streams themselves are choked with muda. As time goes on, the value stream manager can devote more time to individual product considerations—many of them due to changing markets and customer needs. Less time is needed for identifying Current States and achieving Future States once Current States are already performing at a very high level. (At Toyota the core lean knowledge is located in the Operations Management Consulting Division, but the professional staff totals only about sixty for a global organization with $127 billion in revenues.)

  WHEN YOU’VE FIXED SOMETHING , FIX IT AGAIN

  Most managers accept the intellectual proposition that improvement is never finished. Yet we repeatedly visit organizations that make an initial leap to lean and then stop, while talking endlessly about the endless journey. We were therefore delighted recently to revisit Freudenberg-NOK, a firm that had already shown an aptitude for pursuing perfection over an extended period. For example, in Figure 5.1 (page 91), we presented their progress over a three-year period in a vibration damper product line.

  This time we looked at an oil seal product and found a path of steady improvement for a full decade, with no plans to quit. As the diagrams and charts in Figure 15.6 show, FNGP made the initial leap with this product—from process villages to cells—in 1992. (This is the point at which many firms seem to stop.) They followed up this first step with careful attention to the operation of the cells to create “best practice” and then “model” cells in 1993 and 1994. 7 In 1995 they introduced pull systems throughout the facility to send production instructions to the cells and remove products frequently at a fixed pace. In 1998 they undertook a Production Preparation Process (3-P) for a new product generation. And in 2000 they applied all the tools of Six Sigma to improve the capability of their process to a point where scrap is less than one tenth of a percent in an industry where no one else is below 1 percent. Perhaps a TPM program to obtain 100 percent equipment availability is next?

  F IGURE 15.6A: F REUDENBERG -NOK

  F IGURE 15.6B: F REUDENBERG -NOK

  F IGURE 15.6C: F REUDENBERG -NOK

  F IGURE 15.6D: F REUDENBERG -NOK

  F IGURE 15.6E: F REUDENBERG -NOK S UMMARY

  In any case, the point is clear. It really is possible to continue improvements indefinitely for the same value stream. The question is whether value stream managers (and their top-level superiors) will emphatically demand truly continuous improvements and whether the Lean Promotion Office can continually supply the necessary knowledge.

  UTILIZE POLICY DEPLOYMENT

  In the past few years we’ve had extensive experience with policy deployment in our own research institutes. And it’s the hardest thing we’ve tried to do. Policy deployment forces senior managers to make painful choices about what is really most important for the organization and what is truly achievable. At the same time, policy deployment exposes the contradictions between the plans of every unit of the organization as these affect the other units.

  We wish we could say it gets easier. But it doesn’t. Old conflicts will always give way to new in any organization as long as it is growing or faces resource constraints. So the intensity of the polic
y deployment process seems to be a constant. What’s more, we have found that the process can be led only by the senior executive.

  Paradoxically, we have discovered that the actual plans emerging from our policy deployment exercises are only good for about three months, despite our hopes (and our initial expectations) that they would guide our organizations for at least a year. As we reflected on this, we remembered a principle central to lean thinking: A value creation system must be flexible and responsive because forecasts are always wrong. And we realized that a policy deployment plan is nothing more than an organizational forecast, which future events quickly conspire to prove wrong.

  At first we were bewildered but then found that Toyota long ago discovered the same thing. Today Toyota senior managers commonly note that “planning is invaluable but plans are worthless.” 8 Their conclusion is that going through the process forces everyone in the organization to understand the needs and constraints of everyone else and greatly heightens consciousness about the most promising future path even if the specific course of action chosen during the process needs frequent modification .

  CONVINCE YOUR SUPPLIERS AND CUSTOMERS TO TAKE THE STEPS JUST DESCRIBED