Lean Thinking

Home > Other > Lean Thinking > Page 10
Lean Thinking Page 10

by Daniel T Jones


  The first step along the path, beginning in 1989, was to reduce bin sizes and to relocate parts by size and by frequency of demand. Trying to stock or pick a truck fender along with a spark plug on the same run was causing lost parts and the use of grossly oversized equipment, so it was important to segregate parts into small, medium, and large categories with their own sections of the warehouse. As this was done, those parts most frequently demanded were moved closest to the start of the sorting and picking runs and the length of the aisles was reduced markedly. The consequences of these steps for the layout of a typical PDC are shown in Figures 4.1 and 4.2 . Note that a typical picking route was much shorter after downsizing and reorganization of bin locations. However, it’s also important to note that because the batch size of replenishment orders was not changed, the total amount of a given part on hand remained the same. The extra stocks were stored in the “Reserves” area of the warehouse and moved to the “Active” bins as required.

  F IGURE 4.1: T OYOTA PDC B EFORE L EAN T HINKING

  The next step, beginning at the end of 1990, was to introduce the concepts of standard work and visual control by dividing the workday into twelve-minute cycles. An interval of this length was found to be the best compromise between walking distance and cart size in making a round of the bins to load or unload a cart. During each cycle an “associate,” as hourly workers were now called, was expected to pick or bin a different number of lines, depending on the size of the part. For example, in a twelve-minute picking cycle an associate might pick thirty lines of small parts or twenty lines of medium parts or twelve lines of large parts.

  F IGURE 4.2: T OYOTA PDC A FTER D OWNSIZING

  A progress control board was constructed between the receiving dock and the shipping dock to show everyone the number of cycles to be completed and the time available. Each associate was given a stack of magnetic markers of a given color and asked to place a marker in the appropriate square on the progress control board each time a cycle was completed. This made it possible for everyone on the team to see exactly how the work was proceeding, in a striking example of visual control in a warehouse where everyone works out of contact with everyone else. The progress control board eliminated the need for “team leaders,” as the supervisors were now called, to “supervise” their teams. Instead, everyone could look at the board, observe that one worker was falling behind, and provide that worker with a bit of help once other tasks were finished.

  Visual control along with the use of exact work cycles also made it possible to address the causes of disruptions in work flow. The right side of the progress control board provided a blank area beside each cycle for associates to write in the reason that a cycle could not be completed on time. These reasons, when summarized, became the raw materials for directing work team kaizen activities when these were introduced in 1992.

  One of the first kaizen activities was for the teams to build new work carts, using scrap materials and parts from local building supply stores, so the carts were right-sized for each type of picking or binning task. The carts were also designed to hold just the right number of parts—for example, with thirty part holding cubicles for routes for small parts—to provide another form of visual control.

  At the same time the precise picking cycles were being introduced, Toyota’s master computer back in Torrance was being reprogrammed to group orders from dealers by bin location in each PDC so that a set of picking labels in precise bin order was printed out at the beginning of each shift at each PDC. The picking labels were divided into twelve-minute cycles—based on the size of the parts and the knowledge of the team leader about current conditions in the PDC—and placed in pigeonholes in a dispatch box. The pickers obtained their jobs of exactly twelve minutes duration from the dispatch box, always taking labels from the next available slot so there could be no possibility of favoritism in work assignments. In this way, each associate was given five assignments per hour and the work could proceed in a smooth flow from the shelves to the shipping dock. Posting start times above the slots and visually controlling completion times also eliminated another traditional warehousing problem of working ahead to “beat the system.” This practice invariably led to quality problems as associates in their haste picked the wrong part or put parts in the wrong bins.

  After six years of work Toyota was ready in August 1995 to transition from weekly to daily orders from its dealers and to do this without the need for an additional headcount at the PDCs. Indeed, at the end of 1995, the twenty-two pickers at the Toyota PDC near Boston were picking 5,300 lines per day while the hundred pickers at the Chrysler parts warehouse across the road were picking 9,500 lines per day using traditional methods, a productivity difference of 2.5 to 1.

  When the new Toyota Daily Ordering System (TDOS) is combined with the relocation of the PRC for Japanese-sourced parts from Japan to Ontario, California, in October 1996 and the replenishment time to the PDCs from the PRCs is reduced from forty to seven days, it will be possible to dramatically reduce the stocks in the PDCs by eliminating the reserve stocks, as shown in Figure 4.3 . The ability to get parts resupplied very quickly from the next level of the system, and therefore the ability to reorder in small amounts, is always the secret to reducing total inventories in a complex production and supply stream .

  Technology for Lean Distribution

  It is important to note that the Toyota PDCs are dramatically boosting productivity and reducing space requirements without resort to any spending for new technology. Indeed, the company has recently conducted its own test of the most appropriate technology for lean distribution by automating the Chicago PDC while converting the other ten PDCs in accord with the methods just described.

  F IGURE 4.3: T OYOTA PDC A FTER D OWNSIZING , TDOS, AND R APID R EPLENISHMENT FROM PRCs

  The Chicago experiment was undertaken at the end of the 1980s when Toyota back in Japan was obsessed with a shortage of workers during the Bubble Economy and was pressing ahead with much higher levels of assembly automation at its new Tahara plant near Toyota City. It seemed appropriate to try a high level of warehouse automation as well and the objective in Chicago was to completely automate the actual stocking and picking of parts.

  By 1994, after much effort and enormous cost, the Chicago PDC was fully automated but productivity per employee lagged behind the other PDCs implementing standard work, visual control, and efficient bin size and location. While some direct effort was saved in Chicago, the amount of technical support needed to maintain the complex system offset the gains in direct labor and the capital costs made the whole approach uneconomic. We’ll have more to say in Chapter 10 about “appropriate” technology for a lean system and how to select it.

  Level Scheduling Needs Level Selling

  As Toyota thought more about installing a pull system in service parts production and distribution, another benefit emerged. If inventories and handling costs for service and crash parts could be slashed dramatically as the North American suppliers and warehouses implemented lean techniques and if production of more parts could be transferred from high-yen Japan to North America, it should be possible to offer the highest-quality and lowest-cost service and crash parts to Toyota dealers. If this were possible, special promotions to temporarily lower prices and boost sales—the bane of every distribution and production system in every industry—could be eliminated. Toyota dealers would always have the best deal for their customers.

  In 1994, Toyota and its dealers together spent $32 million in the United States on direct mail, print, and broadcast advertising for “specials,” offers by dealers to Toyota owners to perform anything from oil changes to complete maintenance programs at far below the “normal” price. They made these offers because the cost of “genuine” Toyota parts and dealer service was at best equal to—but often much higher—than the customer’s best alternative, the independent garage or mass merchandiser. So promotions were conducted to bring in more service customers for limited periods, partly to suppor
t customer retention, partly in hopes owners could be enticed into looking at new Toyotas while at the dealer to service their current model.

  The problem with promotions was very simple. They required the production of large amounts of parts in advance, yet it was never possible to predict how many would actually be needed. When not all of the parts made were actually needed, dealers shipped them back to the PDC and the PDC temporarily stopped ordering from suppliers until the excess inventory was consumed. Here we see one of the mechanisms of the familiar “pogo stick” phenomenon of “chaotic” orders coming into production facilities when the end market itself is actually quite stable, a tendency we’ll examine further in a moment.

  The net result was a temporary increase in Toyota orders to suppliers to a level far above long-term average demand (in order to build stocks for the promotion), followed by a dramatic drop in orders to far below long-term average demand. This was costly in both directions, requiring overtime in parts plants during the upswing and causing excess capacity during the downswing. It also created costs in the distribution channel to ship excess parts back from the dealers and for the excess stocking and picking costs of running the same parts through the warehouse system twice. The solution was to concentrate on “level selling” by keeping prices constant and making replacement parts at the exact rate parts were being sold. 5

  As Toyota executives thought about applying pull to the entire value stream, from the dealer service bay all the way back to the bumper chromer and similar “second-tier” suppliers, the more advantages they could see. But they knew it would be very hard to persuade the dealers to go along. They come from generations of batch-and-queue thinking.

  The Bad Old Days of Car Service

  Whenever we drive by a car dealer our first thought is always the same: “Look at all that muda, the vast lot of cars already made which no one wants.” Similarly, when we see the large banner out in front offering “rebates” off list prices and “specials” on service and parts, we wonder, “Why did the dealer order cars and service parts which aren’t needed, and why did the factory build cars and parts in advance of customer pull?”

  The answer lies partly in the unresponsiveness of mass-production car makers. Chrysler in the United States is currently trying to reduce the wait for a specially ordered car from sixty-eight to sixteen days, yet for a generation, already, Toyota’s lean production system has been able to build and deliver cars to order in Japan in about a week. Out of fear of losing sales to “impulse purchasers,” mass producers create vast seas of cars on dealer lots, one of practically every specification, so no buyer need walk away unsatisfied. (Converting all factories to flow systems can deal with this problem, as we have already shown.)

  But the answer also lies in the mentality of retailers and customers across the world. Dealers love to “deal” and the public loves a “sale.” (One of us took a trip to France some years ago and discovered that the only phrase of high school French our wife could remember was “on sale”!) Changing the way retailers and consumers think about the process of ordering goods and making transactions may be difficult, but as we will see, it is essential to doing things a better way.

  Pulling from the Service Bay

  Most readers, we hope, have never been into the parts storage area of a car dealership. It’s generally a horrible sight. When we first visited the Parts Department at Bob Sloane’s Toyota near Philadelphia in 1994, we found a rabbit warren of rickety shelves, meandering aisles, and dim lighting in two separate buildings. Clearly, the physical flow of parts was an orphan activity compared with the income-producing service bays for car repairs and the showroom where cars are sold.

  When we first visited Sloane Toyota, the dealership had about a three-month supply of the average part, creating an inventory of about $580,000 in service and crash parts. When a car was driven into Sloane for repairs, it was taken to a service bay where the technician evaluated the problem and determined what parts would be needed. The technician then went to the parts window, requested the necessary parts, and waited for the counter person to go and get them somewhere in the labyrinth of bins and aisles.

  Because Sloane received most parts in weekly batches, the workload of the parts stockers who took the parts from the receiving area and put them in the proper bin was very erratic. It generally took three days to get all of the parts from the receiving area into the bins, with the result that the counter person would often find empty bins when the computer showed the parts were in stock. Indeed they were, but they were “missing in action” somewhere between the receiving area and the proper bin. This knowledge touched off “treasure hunts,” which are the distribution equivalent of the expediting always necessary in batch-and-queue production operations. Good counter staff would generally find their part, but the whole exercise was inherently wasteful. (The highly skilled technician, meanwhile, was standing idly at the window all the time the counter person was treasure-hunting.)

  In 1995, when Sloane Toyota joined Toyota’s campaign to introduce pull in the whole parts distribution and manufacturing system, it reorganized its parts storage area just the way Toyota reorganized its PDCs. By cutting the size of the bins dramatically, generally by three quarters, and reorganizing all parts storage into one building, Sloane found it possible to increase part numbers on hand by 25 percent (including Bob Scott’s bumper) while cutting its storage area in half and reducing its parts inventory from $580,000 to $290,000. While freeing up $290,000 in cash from inventory Sloane was able to add four new revenue-producing service bays, created with practically no capital investment, in the empty second parts warehouse.

  Sloane Toyota found that the number of cars which could receive “same-day service” went up substantially (reducing the number of cars in its overnight “loaner” fleet) even as its inventory generated cash and the number of parts the average picker could gather in a given period of time more than doubled. Most important, customers were happier because their cars were more likely to get fixed right away and the total cost of service had fallen dramatically. Indeed, Bob Scott was able to get his truck’s bumper replaced the same day .

  Pulling from Service Bay to Raw Materials

  We can see the full magnitude of what is happening by “pulling” together all the pieces of the service value stream. By the end of 1996, when Toyota’s new pull system will be in place throughout North America, the request of the customer arriving in a Toyota dealer service bay will become the trigger for pulling parts through four replenishment loops going all the way back to steel blanks, as shown in Figure 4.4 .

  F IGURE 4.4: P ULL T HROUGH F OUR L OOPS

  Toyota dealers and parts suppliers will still rely on Toyota’s computerized macroforecast for capacity planning to answer questions about the size of manufacturing plants and the number of warehouses that would be needed in the future. However, day-to-day part replenishment will now be handled in a radically different way: Each time a customer requests a part at the service bay, a series of replenishment loops will result eventually in more parts being made by the supplier in a situation which might be called “sell one; buy one” or “ship one; make one.”

  To see what this means, let’s follow the bumper example all the way through the value stream. Before lean techniques were applied to any aspect of the system—that is, prior to 1989—the elapsed time from the arrival of steel blanks at Bumper Works until the bumper made from those blanks was actually installed on a truck was nearly eleven months. Four weeks in Bumper Works, two weeks at Chrome Craft, a few days at the Toledo PRC, six months at the PDC, and three months in Bob Sloane’s parts inventory. (Lead time of this magnitude was the norm, not the exception, for the entire automotive parts industry in North America.)

  By the end of 1995, the elapsed time had fallen to four months: forty-eight hours in Bumper Works and Chrome Craft, a few days in the Toledo warehouse, two months in the PDC, and one and a half months in Bob Sloane’s inventory. And by the end of 1996, elapsed time s
hould fall further to about 2.5 months as both the PDC and Bob Sloane shrink their inventories in response to falling resupply times. At the same time, the percentage of vehicles fixed the same day is increasing substantially, and costs—inventory, warehouse space, and direct labor—are falling dramatically.

  Note that practically no capital equipment has been required. The tool modifications to permit quick changeovers and the specialized stocking carts in the factories and warehouses were created by production workers as part of kaizen activities, and the elaborate MRP systems formerly regulating activities inside the Bumper Works and Chrome Craft plants are no longer needed.

  Just the Beginning

  The savings we describe are just the beginning. Sloane Toyota, Toyota Motor Sales, Bumper Works, and Chrome Craft are now working on the value stream for service and crash parts as a lean enterprise under Toyota’s leadership and are deeply committed to the concept of perfection, which we will discuss in the next chapter. They all expect to steadily reduce the elapsed time and cost of service parts. (Superlative quality is taken as a given, but quality will improve as well, as a natural complement to flow and pull.) One approach will be to extend the smooth-flowing value stream all the way to raw materials by helping the steel maker and steel fabricator overcome their current batch-and-queue thinking. At the other end of the stream, with encouragement and help from the dealer, customers may be able to schedule many of their service requirements in advance so the need for parts can be precisely predicted.

  The parent Toyota company began to pursue this latter approach in Japan shortly after the 1982 merger of Toyota Motor Sales and the Toyota Motor Company, which formed the current-day Toyota Motor Corporation. Between 1982 and 1990, Toyota reorganized its service and crash parts business in a manner identical to the new North American pattern, except that it took two additional steps. It created Local Distribution Centers (LDCs) in each metropolitan area (jointly owned with the dealers) and took practically all of the parts stock out of dealerships with the result that Toyota dealers in Japan now carry only a three-day supply of forty commodity parts like windshield wiper blades. It then encouraged dealers to work intensively with every customer to preschedule maintenance so that parts needs could be precisely predicted in advance.

 

‹ Prev