Lean Thinking

by Daniel T Jones

  Ron Hicks was proposing a new system of standard work and takt time that sounded like oversight by the industrial engineer, which every skilled tradesman hates. (The difference, of course, was that the work team would standardize its own work.) What was more, he was proposing making complete machines one at a time. Finally, he claimed that if the work was standardized by the work team, the machines were realigned to permit single-piece flow, and takt time was adhered to with no working ahead, there would be no more fires to fight. As Underwood remembers, “It didn’t sound like much fun and I thought it would never work.”

  When the conversion week was completed and the new cell was ready to go … it didn’t work. All kinds of problems, long submerged in Lantech’s massive inventories and closely held work practices, suddenly emerged. Some steps had not been included in the standard work charts. Poor tool maintenance—easily tolerated in the old batch system—repeatedly stopped the whole cell. The supply of components to the cell was not dependable. The widespread feeling was that Ron Hicks was pushing a novel concept that would never work at Lantech.

  At this point Jose Zabaneh, the production manager, played the key role: “I was so fed up with our failures and so taken with the logic of the new system that I threw my heart into it. I called a meeting of the workforce and announced that I would stay all night and all weekend to work hands-on on fixes to the problems we were encountering with the new cell, but that I would not spend one second discussing the possibility of going back to the old batch-and-queue system.”

  Pat Lancaster gave unfaltering support to the new system, Ron Hicks (along with his consultant, Anand Sharma, who had advised him earlier on the conversion of Hennessey) had the technical skills to work the bugs out, and Jose Zabaneh was “our spark plug.” Gradually, it all began to come together.

  (We’ll see that these three attributes—taking the long view, technical virtuosity, and a passionate will to succeed—are essential for any organization making the lean transition. Sometimes they are possessed by a single individual, sometimes, as at Lantech, they are shared by a group of leaders. However initially distributed, they are all necessary and eventually they must be shared by the whole organization.)

  By the fall of 1992, the whole Lantech production system had been converted from batches to single-piece flow, including the cell for the largest Lantech machine—the $50,000 H model—made at the rate of only one per week. The plant now looked as shown in Figure 6.5 .

  The consequences for performance were truly staggering. Although Lantech’s headcount stayed constant at three hundred, the number of machines shipped doubled between 1991 and 1995. (The sales growth was due to a general recovery in the market, aggressive pricing by Lantech to capture share, and a host of new products, to be described in a moment.) The plant, which had been bulging at the seams with inventory, now had 30 percent excess space despite the doubling of output. The number of defects reported by customers fell from 8 per machine in 1991 to .8 per machine in 1995. Production throughput time, as we have noted, fell from sixteen weeks to fourteen hours. The percentage of machines shipped on the date agreed with the customer went from 20 to 90 percent.

  To speed this remarkable transition, Pat Lancaster made two promises to his workforce. These seemed almost quixotic in 1992, given the financial condition of the firm, but have proved critical to success. First, he promised that no one would be let go because of the lean conversion. Instead, a kaizen team was created from freed-up workers who were deployed to plan the improvement of other activities. Bob Underwood, the original skeptic and chief “fireman,” was made head of this team. After every improvement, the best (not the worst) workers in the revamped process are transferred to the kaizen team, making clear that this is a promotion, not a punishment. The steady growth in output of the newly competitive Lantech has meant that within a short period these workers have been needed again for production work.

  F IGURE 6.5: N EW L ANTECH P RODUCTION F LOW

  At the same time, Lancaster reviewed Lantech’s wage policy and adjusted the base wage upward from about $7.00 to about $8.50 per hour. As Ron Hicks noted, “We had been running unskilled employees through like McDonald’s, with a sharp premium for our small core of skilled workers. It quickly became apparent that all workers in the new Lantech would be skilled workers, but with a very different type of skills. So we had to pay all of them a better wage. As a result, turnover quickly fell to just about zero.” (Note that because each machine is now being made with one half of the formerly needed hours of human effort, a 25 percent wage increase is easily affordable.)

  • • •

  As the lean revolution gained momentum in the plant, it was time to turn to the office and in particular the order-taking process. As Pat Lancaster put it, “We wanted the goodness of the plant to suck the badness out of the office. If we could make a machine in fourteen hours, how could we live with an order-taking process which required three weeks?” In one notable case, Lantech made and delivered a machine in four days—long before the credit check could be completed—only to discover that the buyer was insolvent.

  The technique employed to transform the office was exactly the same. Lantech set up a kaizen team to collectively rethink the process. It included all the workers involved in a specific process, the firm’s technical experts—including production workers from the plant kaizen team and one outside consultant (Sharma). The group mapped the entire value flow and looked for wasted time and effort. As each process was rethought and turned from batch-and-queue into flow, the best of the workforce was assigned to the kaizen team to lay the groundwork for the next process review. No one was laid off and the move to the kaizen team was clearly an acknowledgment of superior performance.

  When these techniques were applied to the entire order-taking and plant-scheduling system, the results were truly astounding. Because Lantech now understood its costs much better, it was possible to publish fixed prices on all but truly custom-built machines and to eliminate the haggling step between Lantech and the distributor. The order itself, once at Lantech, could be inserted into the production schedule in only two days.

  Perhaps most remarkably, most of the computerized scheduling system was no longer needed. MRP was retained for long-term materials ordering from suppliers, but day-to-day scheduling is now run off a large white board in the sales office. The production day is divided into slots by takt times and orders are written on the board as they are confirmed. At the times we have visited Lantech, the white board slots have been filled anywhere from three days to two weeks ahead of the current date and no machine will be made except in response to a confirmed order.

  A large white board easily visible to everyone in the firm has proved to be a remarkable spur to the sales force, particularly as the amount of filled space gets smaller and the amount of empty space gets larger. This is an excellent example of yet another lean technique, visual control, in which the status of an activity is displayed so every employee can see it and take appropriate action.

  The final step in this process is to copy down each evening the roster of machines to be made the next day and to take this list to the four production cells. For each machine, the cell is given the name of the actual customer and the promised delivery date, typically two days from the start of the build sequence for high-volume machines and ten days for the lowest-volume, large machine. The former Management Information System Department with its seven full-time workers has been eliminated because the parts within the plant are pulled along to the next workstation automatically. Information flows that had been automated have now been completely eliminated because product and information have been combined into one. The full results, as shown in Figure 6.6 , can be contrasted with the labyrinthine order process shown in Figure 6.2 .

  F IGURE 6.6: N EW L ANTECH O RDER F LOW

  The main transitional problem has been that the distributors and buyers of industrial equipment are unaccustomed to getting rapid and on-schedule deliveries. Orders ha
ve often been “guesstimated” on the presumption that many weeks were available to firm up the precise specification, notify the manufacturer of changes, and plan for the machine’s installation. In one notable case, Lantech made and delivered a machine within one week of the order, just as promised, to find the customer quite upset: “You’ve sent us our machine before we’ve given any thought to how to use it. We thought we were placing an order simply to guarantee ourselves a place in the production line, that we would have time to respecify the options, and that you would deliver late as usual. Now, you’ve gone and made it already!”

  The final step in transforming Lantech has been to rethink the product development process. Pat Lancaster knew from the early days of the plant conversion that he would need to grow his business dramatically in order to keep everyone busy, as he promised, while productivity zoomed. This meant turning strategic thinking on its head: “I didn’t have time to find a brandnew business to go into and I didn’t have the money to buy out any of my major competitors. Instead, I needed to revitalize and expand my product range so I could sell more in an established market I knew well. At the same time, I knew that a total redesign of my products to make manufacturability a key consideration would slash my costs even further and dramatically improve quality and flexibility for the customer.”

  He also knew that his batch-and-queue product development system would take years to come up with market-expanding products if not given the same treatment as the plant and the office. He wanted to put new product designs into single-piece flow, just like orders and machines. “We needed the design to move continuously from the initial concept to the launch of production. This meant no stopping due to the bureaucratic needs of our organization, no backflows to correct mistakes, and no hitches during production ramp-up.”

  Lantech had experimented with development teams in the late 1980s and early 1990s but without much success. A few “bet the company” projects were pushed through by a designated “dictator” who was effectively a new type of expediter (slowing down all other projects to get his project through). Otherwise, weak “team leaders” tried to coordinate the activities of the numerous technical specialists needed to develop a complete product, each with their own priority list. In no case was the team leader—dictator or weak coordinator—responsible for the end results of the project: Did the product please the customer and make money for Lantech during its production life? No one was really in charge and not much had changed despite the new “team” terminology.

  In 1993, Lantech went to a new system of dedicated teams led by a Directly Responsible Individual (DRI) clearly charged with the success of the product during its lifetime. The corporate annual planning process identified the major projects to be developed that year and ranked them. A team of dedicated specialists was designated for the two top-ranking projects. This consisted of marketing, mechanical engineering, electrical engineering, manufacturing engineering, purchasing, and production (including hourly workers from the plant kaizen team who would actually build the machine once launched). These teams were co-located and told to work nonstop on the designated project and to do nothing else until it was done. The welter of minor projects which formerly cluttered up the Engineering Department were simply dropped (or “deselected” in Lantech-speak). As the engineering director noted, “We never would have finished them anyway!”

  A kaizen of Lantech’s prototyping process showed that if all of the needed skills were available, a working prototype for the top-ranked project could be put together in a week, a process which formerly would have taken three months. And the presence of the actual production staff on the team quickly identified manufacturing problems which the mechanical and electrical engineers had never imagined.

  The major objections to dedicated teams—that work flow is uneven, so some team members will be underutilized some of the time and teams will be in conflict for scarce skills needed at specific points in development—were overcome in two ways. First, it developed that team members actually had much broader skills than they had ever been asked to use. (After all, they had been reengineering each other’s designs in secret for years!) They could quickly develop additional, narrow skills to address specific problems. Mechanical engineers could actually help manufacturing engineers with their work and the reverse. This meant that the uneven work flow problem could be largely corrected within the team.

  Second, it turned out that a bit of careful scheduling could identify conflicts in requirements for skilled personnel far ahead of time. Moving a few specialists from one team to the other and back, as needed, could solve the problem.

  Under the new Lantech product development system, the progression of the design looks as shown in Figure 6.7 , in contrast with the maze shown in Figure 6.3 :

  F IGURE 6.7: N EW L ANTECH P RODUCT D ESIGN F LOW

  The first product to come through the new system showed its dramatic potential. The new S series, launched in mid-1994, was developed in one year (compared with four years for its predecessor) with about half the effort previously thought to be required. (Remember: There were no delays for lack of personnel or queue time, no backtracking, and no secret rework.) Then, the launch was much smoother than in the past and the number of defects reported by customers was a tiny fraction of the rate experienced with previous new products.

  The Bottom Line

  The conversion of Lantech from classic batch-and-queue to lean thinking has produced a dramatic box score of performance improvements (see Table 6.1 ).

  T ABLE 6.1: T HE L EAN T RANSFORMATION AT L ANTECH

  However, the result by which any business in a market economy must be measured is the ability to make enough profit to renew itself. If the transition at Lantech cost a fortune in new investment or disrupted the firm’s ability to satisfy customers it would be an interesting technical exercise rather than a revolution in business practice.

  In fact, the amount of investment required was substantially zero. The tools were moved and reconfigured, for the most part, by workers freed up from inefficient production tasks. The reconfiguration of the office and the development process were performed much the same way. Fewer computers, less space, and less expensive tooling were needed at every step. And the effect on customers was dramatic: Lantech’s share of the stretch-wrapping market zoomed from 38 percent in 1991 to 50 percent in 1994. As a result, the large operating losses of 1991 were turned into solid profits by 1993 and an industry-leading financial performance by 1994 .

  Work as “Flow”

  As noted in Chapter 3 , the rethinking of work in accord with lean principles produces the potential for greatly expanded experiences of psychological “flow.” Workers in the Lantech manufacturing cells can now see the entire work flow from raw materials to completed machine. Takt time, standard work, and visual control (including posted work charts for all tasks) give an immediate sense of how the work is proceeding. Multiskilling and job rotation make full use of each worker’s skills and the frequent repetition of kaizen events (as described in Chapter 5 on perfection) gives an opportunity to participate actively in work design. The constant elimination of muda and the movement of workers out of work cells as more efficient methods are discovered mean that the work is a constant challenge. Finally, there are few interruptions in the form of line stoppages and sudden demands to shift to a completely different task to deal with a crisis.

  The situation in the office is very similar. Visual control in order-taking makes it clear to everyone where Lantech stands and the new order entry system in which one employee can perform the whole task makes it possible to get immediate results. The kaizen process in the office melds thinking and doing, planning and acting, just as it does in the plant.

  Finally, the rethinking of development work gives a true sense of feedback as everyone involved in a project works in the same space and projects move rapidly to completion. (Formerly, the majority of Lantech’s development activities were never completed because market conditions changed
before the cumbersome development process could be concluded. We have found this same phenomenon in a wide range of firms over many years.) Employees respond positively to gaining new skills and being encouraged to use all the skills they’ve always had. The lack of interruptions and conflict over which task to work on next has come as a great relief.

  As Bob Underwood characterizes the situation now compared to the recent past, “We were living in darkness and now we have come into the light.”

  Yet it would be inaccurate to characterize Lantech as some sort of paradise. Indeed, coming out into the light can be painful to your eyes. The reorganization of work tasks into a continuous flow seems to have produced widespread psychological satisfaction in daily tasks, but it is also producing the need for constant change. “We just get something working smoothly when it’s time to improve it again” is a common refrain, and it’s clear that each change, at least subliminally, carries risks: “Will Lantech really honor its commitment to retain excess workers? Will my contribution to improvement activities be recognized and rewarded?” Perhaps most important, many employees ask, “What will change mean for my career? Am I going anywhere or just flying a holding pattern while Lantech prospers?”

  These are all important questions which firms must face once they make the initial leap to lean thinking. We will return to them in Part III on the challenge of building a lean enterprise.

  The Last Step

  One last step in the conversion of Lantech from a batch-and-queue to a flow organization remains to be discussed. In April 1995, Pat Lancaster promoted himself to the new position of chairman (at age fifty-two) and stepped down from day-to-day operations, turning the CEO job over to his son, Jim. Now he is starting a new creative process by thinking again about the value of his products to the customer.