use lean lot sizing to prepare for economic recovery

Transcription

1 ANtiCiPAtiNg the reemergence of demand

2 use lean lot sizing to prepare for economic recovery By Steve CiMorelli, CfPiM Over the past 12 to 18 months, much has been written about manufacturers cutting inventories to the bone. Production output has lagged behind even the depressed sales levels. A consensus is emerging that we are at or near the bottom of this recession. As a result, production levels must rise to meet current demand because inventory buffers are depleted. Furthermore, recovery may not restore previous demand patterns; meaning, manufacturers may have to deal with a significantly different product mix than before the recession. This all points to a golden opportunity: Incorporating lean manufacturing processes now will enable smaller lot sizes in order to improve agility, lower costs, enhance capacity usage, and reduce inventories. There remains, however, an ongoing debate over manufacturing lot sizing. Lean proponents push for lot sizes of one, or at least for setups measured in single-minutes with correspondingly small lot sizes. On the other hand, many operations management professionals look to minimize total cost by producing an economic run quantity, arguing that this approach maximizes capacity. The truth is that both sides are correct. Manufacturing must run efficiently, taking into consideration the realities of current-state setup times, particularly in capacityconstrained environments where reducing lot sizes without reducing setup times consumes more capacity. The disconnect between these two opinions occurs when current setup times are assumed to be fixed. The solution is to focus lean efforts on setup reduction, moving the operation in the direction of lean. As setups are reduced, corresponding lot size reductions can be implemented, capturing the benefits in speed, agility, and inventory reduction. This may be seen as a cyclical process; each incremental reduction in setup leads to the next reduction in lot size until the business goal has been attained. in the direction of lean A total cost approach, which is the basis for the traditional economic order quantity (EOQ), can provide a target range for establishing current lot sizes. Lean advocates fight strenuously against the EOQ and they have valid arguments. They rightly say, for example, that the EOQ understates the true cost of carrying inventory and that costs such as obsolescence, handling, damage, floor space, and insurance are not considered. This objection can be overcome easily by increasing the carrying cost rate from a company s standard value (typically in the 10 to 12 percent range) to a value that reflects the missing costs say, 20 to 25 percent. A second argument is that the underlying cost curves are not linear, as assumed by the EOQ formula, but are more like step functions. Again, this can be surmounted by recognizing that the true economic lot size may not be the precise EOQ value, but something in a plus-orminus range of that value. A quick glance at the total cost curve shows that the EOQ exists in a range where the curve is relatively flat. (See Figure 1.) Therefore, any lot size in the vicinity of the EOQ achieves the objective of minimizing total cost. This enables the actual lot size to be rounded to a container quantity or some other value that makes sense based upon the manufacturing process, material usage constraints, shift scheduling issues, and the like. In each following example, A represents annual usage, S is the setup cost, C is the unit cost per part, and I is the inventory carrying cost expressed as an annual percentage rate. The goal is to constantly move in the direction of lean, giving a bias toward rounding down the EOQ. Management s role is to encourage, incentivize, or even direct this bias. The example in Figure 1 shows that the total cost is minimized within a range running approximately 200 to 500 pieces. The EOQ of 302 pieces can be rounded anywhere within this range. This example uses a carrying cost of 11 percent. If that value is increased to 20 percent to better capture the full cost of inventory, the EOQ shifts to a lower value of 224 pieces, as shown in Figure 2. In other words, it has moved in the direction of lean. An effective compromise between the two camps can be achieved as follows: 1. Increase the company standard carrying cost to consider the full cost of inventory. 2. Establish current lot sizes near the EOQ, with a bias to round down toward the left-most edge of the flat portion of the total cost curve. 3. Work aggressively to reduce setup times. 4. Follow each setup reduction with a review and recalculation of lot sizes. Reset lot sizes to capture the incremental benefit of each setup improvement. Continuing with the example, if setup is reduced by 50 percent, the impact can be seen in Figure 3. APICS magazine January/February

3 FIGURE 1: total ANNuAl CoSt, example 1 FIGURE 2: total ANNuAl CoSt, example 2 FIGURE 3: total ANNuAl CoSt, example 3 Again, the EOQ has moved down, to a value of 158 pieces, continuing the progression in a lean direction. The total cost curve indicates that a lot size as low as 100 would be reasonable and would significantly minimize total cost. AN ideal SCeNArio The immediate benefit of setup reduction is achieving economical production of much smaller lot sizes without negatively affecting capacity or cost. Smaller lot sizes in turn enable pull processes, such as kanban, to facilitate production prioritization and scheduling much closer to actual demand and consumption from downstream customers. This enhances the agility of the entire operation. Inventory buffers become smaller or may disappear completely as the operation approaches the ideal of single-piece flow (lot sizes of one). Consider a scenario at a manufacturing plant in which a fabrication department produces several hundred different component parts for in-house assembly operations. Setup times on these machines are in the neighborhood of 90 minutes, with setup defined as the last piece of a prior production run to the first good piece of the new production run. Demand is well within available capacity, with forecasts indicating modest growth and a significantly varied product mix. Because of the long setups and large production runs, each part is being worked on for a week or more, creating long queue times for open production orders. This leads to higher 40 January/February 2010 APICS magazine

4 inventory levels to buffer assembly from the long wait times out of the fabrication shop. The key opportunity for this plant s operation is to reduce lot sizes and become more agile, making it possible to respond quickly to emerging demand patterns instead of enduring long wait and queue times between jobs. Figure 4 shows an initial available annual capacity of 1,832 hours, after accounting for maintenance downtime. Figure 5 shows the current-state EOQs, calculated from initial setup values, and the corresponding capacity load each product requires. For example, part A has a demand forecast of 45,000 pieces and an EOQ of 17,930. Because of the large EOQ (resulting from the high setup cost), the entire annual demand has produced two-to-three large batch runs, each resulting in nearly five months of inventory. Recalling that average inventory equals safety stock plus one-half lot size and assuming zero safety stock the average inventory on hand for this part would be one-half EOQ, or 8,965 pieces. Repeating these calculations for the remaining parts and summarizing the results indicates a total capacity load of 1,644 hours and an average inventory of 25,775 pieces. Notice that the capacity needed to produce these 5 parts (1,644 hours per year) is well within the available capacity of 1,832 hours. The inventory investment is quite high, with parts being produced in the range of two-to-four times per year in quantities that represent up to five months of supply. Consider now the impact on performance if setup times are reduced from 90 minutes to 15 minutes (see Figure 6). With all other assumptions held constant, the EOQs fall dramatically, with the number of setups per year increasing. Because each setup now takes much less time, the total annual load remains nearly unchanged at 1,630 hours a slight decrease and still well within the 1,832 hours available. Note also the dramatically improved inventory investment, down from 25,775 pieces to 10,525 pieces. Rounding down each EOQ to a standard packaging quantity would further improve these numbers. The net result is a far more agile and responsive operation, capable of meeting all demand placed on it, while still maintaining available capacity for projected growth. Lead times to internal customers will shrink by corresponding amounts because time spent waiting in queue and physically running each batch will be dramatically less, likely leading to further inventory reductions. A counter argument Returning for a moment to lean arguments against the EOQ, consider what happens if lot sizes are cut without first reducing setup times and costs. Figure 7 shows the impact of reducing lot sizes to one-fourth the original EOQ, with setup times remaining at 90 minutes ($75 per setup). FIGURE 4: initial available capacity Capacity variables Hours per year 1 Shift (40 hours per week) 2080 Weekly maintenance (4 hours per week) 208 Annual maintenance 40 Available capacity 1832 FIGURE 5: initial capacity load and average inventory values EOQ variables Setup time minutes: 90 Setup cost: ($/hour) $50 Setup cost: (S) $75 Production cost ($/hour): $25 Inventory carrying cost: (I) 20% In the original case, the total number of setups was 16. At $75 per setup, this equates to an annual setup cost of $1,200. In the new scenario, 64 setups are required, equating to an annual setup cost of $4,800. Notice also that the annual load has grown Calculations: 1. Setup cost = 1.5 hours x $50/hour = $75 (assumed fixed for each part) 2. Cost/part = run time x $25/hour 3. EOQ = Sqrt(2AS/IC) 4. Annual load = sum of S/U + (run x EOQ) Part Demand Run time Cost per EOQ Setups/ Load forecast (A) (hours) part (C) year (hours/year) A 45, $ , B 12, $0.42 4, C 60, $ , D 35, $0.33 8, E 16, $0.42 5, Total load 1,644 Average inventory (1/2 EOQ): 25,775 FIGURE 6: revised capacity load and average inventory values New EOQ variables New setup time (minutes): 15 New setup cost: $12.50 Part Demand Run time Cost per EOQ Setups/ Load forecast (A) (hours) part (C) year (hours/year) A 45, $0.11 7, B 12, $0.42 1, C 60, $0.21 6, D 35, $0.33 3, E 16, $0.42 2, Total load 1,630 Average inventory (1/2 EOQ): 10,525 FIGURE 7: forced lot size reduction Part Demand Run time Cost per Lot size Setups/ Load forecast (A) (hours) part (C) (1/4 EOQ) year (hours/year) A 45, $0.11 4, B 12, $0.42 1, C 60, $0.21 3, D 35, $0.33 2, E 16, $0.42 1, Total load 1,717 Average inventory (1/2 EOQ): 6,444 to 1,717 hours a significant increase, but still well within available capacity. This scenario would not be appropriate if the new load exceeded capacity. In that case, the cost of adding capacity would have to be considered before any final decisions were made. APICS magazine January/February

5 Finally, notice the much lower average inventory investment of 6,444 pieces versus 25,775 in the original scenario. Figure 8 summarizes the results of the three scenarios. This begs the following guestions: Is the increased setup cost real, or are the resources used to perform the setups actually sunk costs (expenses that already have been incurred and cannot be recovered)? Even if the additional $3,600 in setup costs is valid, are the improvements in agility and reduced inventory worth that investment? Keep in mind that, while EOQ attempts to minimize the total life FIGURE 8: SuMMAry of CAPACity, load, ANd inventory ,000 25,000 20,000 15,000 10,000 5,000 0 Original EOQ EOQ2 1/4 original EOQ Capacity (hours) Annual load (hours) Inventory (pieces) cycle cost of inventory, it does not address the value of improved speed and agility. Lot size reduction offers significant improvements in operational speed and agility, characteristics that will position manufacturers to rapidly meet restored demand from existing customers and capture new business via an improved competitive posture. The examples illustrated here show the value of smaller lot sizes, and after all the pain and struggles associated with managing inventory reduction over the past months, most operations managers will be loath to add inventory back unnecessarily. Manufacturers who pursue setup and lot size reduction now will find they have a competitive advantage for meeting whatever new demand patterns emerge during economic recovery. Steve Cimorelli, CFPIM, is president of SCC Inventory Consulting LLC. He may be contacted at (321) or steve.cimorelli@ sccinventory.com. To comment on this article, send a message to feedback@apics.org. APICS Extra Live: Increasing Competitive Advantage and Agility through Lean Operations Presented by: Steve Cimorelli Date: February 16, 2009 Time: 1:00 p.m.-2:00 p.m. CT Attend APICS Extra Live to gain deeper insight into Steve Cimorelli s January/February APICS magazine article, which demonstrates how lean lot sizing can help a company be more agile. In this APICS Extra Live, learn how lean operations can increase a company s agility and competitive advantage through an effective sales and operations planning foundation, pull systems, and a sound prioritization scheme to initiate pull based on actual customer demand. DON T MISS APICS EXTRA LIVE Presented by: Steve Cimorelli Date: February 16, 2009 Time: 1:00 p.m. 2:00 p.m. CT REGISTER ONLINE AT APICS.ORG/EXTRA. 42 January/February 2010 APICS magazine