N avigating Troubled Waters 42 July/August 2010 APICS magazine A voyage to more stable safety stock and service levels By Steve Johanson
m any operations management professionals treat working capital and service-level targets as two separate key performance indi- cators (KPIs); in fact, they are different sides of the same equation. There is a rigid relationship between the two and proven statistical curves that define their relationship. However, it is more common than not that safety stock is adjusted by past experience and rules of thumb. In such an environment, safety stock becomes quite like an ocean tide: Eventually, it will flood. APICS magazine July/August 2010 43
W ith statistically calculated safety stocks, an operations manager finally can take a proactive stance. When managers bump up target values in response to crisis, it creates swells of working capital. Cash-flow concerns shift toward inventory reduction, and the wave begins to recede as organizations embark on initiatives to reduce and rationalize inventory. Of course, shortages soon arise, and the floods come once more. Constant changes—such as production adjustments, modifications to distribution and manufacturing assets, and process-improvement projects—keep the cycle in motion. In order to optimize the balance between service level and safety stock, professionals must constantly calculate valid, statistical safety stock targets. Most enterprise resources planning (ERP) systems perform a safety stock calculation. But very few include in the system all sources of variability as inputs to the safety stock formula. Furthermore, ERP tools rarely calculate accurate safety stock inputs or correct erroneous data. Figure 1 shows 13 basic safety stock inputs. They include operational decisions, performance measures, constraints, and environmental factors. All play a pivotal role, and only extensive sensitivity analysis can make it possible to discount any one factor. Omission of a variable is the first major problem with setting safety stocks. Miscalculation is the second. This is clear from a situation that occurred at a sausage manufacturer. Demand was smooth, forecasts were good, and safety stocks were greater than calculated targets. However, employees were fighting what appeared to be uncontrollable system shortages. On closer examination, the calculations excluded manufacturing time from lead time. (Smoking and curing averaged 20 days.) Plus, systems discounted manufacturing time variability. (For instance, sausage cures at plus-or-minus six days.) While original safety stock estimates were based on short, invariable delivery times and good forecasting, the correct target was significantly more than both what was calculated previously and what managers thought was a surplus. Note that forecast error is treated as an environmental factor. Often, time shortages are blamed on bad forecasting. But rarely does a company change its forecast error—meaning, past forecast errors are very good estimates of future ones. Look at your own forecast error over time for any collection of stockkeeping units (SKUs) and distribution centers, and you will see the error is very constant. Figure 1: Safety stock inputs Planning lead time Supply chain decisions Cycle time Transit mode/time Manufacturing time Management time variability Statistical safety stock Supply chain performance Transit time variability Manufacturing schedule adherence Transit schedule adherence Demand environment Forecast error Demand variability Service level Requirements and restraints Ship life Capacity loading 44 July/August 2010 APICS magazine
Figure 2: Inventory reduction roadmap Current state Reduce cycle lengths/stabilize planning Cut manufacturing planning lead time Achieve 95 percent schedule adherence Reduce lot release time Improve forecast error Further cycle length reductions 0 5 10 15 20 Inventory turns Capacity loading often is treated as a deterministic constraint. You have a fixed amount of capacity, and it does not vary. In reality, it is a stochastic constraint and plays into the balancing act between safety stock and service level. This means that, the closer a work center is loaded to 100 percent, the more safety stock will be required. This stock is used to counter the risk of manufacturing possibly failing to replenish 100 percent of what was sold in any given cycle. There are several approaches to correcting the quandary of safety stock versus service level. First, operations managers need to audit their own processes by asking four questions: 1. Do we capture all 13 factors, constraints, and variables in the system? 2. Are they measured properly? 3. Are we filtering erroneous data points from formulas? 4. Are we calculating factors for every node? Second, a process must be established that acts upon these requirements. Most importantly, the cycle needs to occur at the same frequency as periodic planning processes. If master planning is done quarterly, set safety stocks quarterly. This might include redefining business processes and incorporating a tool that can support all the data and the periodic workflow. A The good news is that most conventional ERP solutions are correct. The bad news is that this only is true for the variables being considered. If forecast error is taken into account as an input, the result will be correct only if all other variables are zero. However, if all 13 are included and the data are clean, then the safety stock number likely is accurate. With statistically calculated safety stocks, an operations manager finally can take a proactive stance. Armed with a robust model, it’s possible to develop a plan that explicitly lays out the requirements to reduce inventories. Figure 2 is a roadmap made by a food manufacturer, which illustrates the steps required to achieve an inventory turns target. The approach focuses on improvement, quantifies that improvement in terms of cost, and identifies responsibility for the improvement. Without a complete statistical model, this is impossible. In addition, it puts the business at risk of tipping the balance toward high turns and low service levels. There are two approaches, the results of which rarely drive up total landed costs elsewhere. They are using a hybrid deployment plan by mixing one-tier and two-tier distribution achieving flexible service levels by having different service-level targets by SKU location. facility that follows only one scheme or the other for all of its SKUs relinquishes considerable savings. APICS magazine July/August 2010 45
To comment on this article, send a message to firstname.lastname@example.org. These solutions are independent of cost considerations that can be made in isolation. They also are straightforward if one already is calculating a good statistical safety stock. Hybrid deployment means that some SKUs get stocked in a buffer or plant warehouse that service a distribution network (tier 2), while others are deployed directly to the distribution centers when they are made (tier 1). When looking at a complete product portfolio, it is unusual for the most cost-effective scenario to involve all SKUs either in tier 1 or tier 2. A facility that follows only one scheme or the other for all of its SKUs relinquishes considerable savings. Having flexible service levels means that, within a given set of products and locations, planners won’t have to try to achieve a single service-level target for every location and every SKU. The effect is that easy-to-serve nodes (high volume and low forecast error) supplement more difficult ones (low volume and high forecast error). Like hybrid deployment, these calculations are done easily with a good statistical model. It is the execution that proves difficult. Finally, to successfully balance service levels and inventory, one must adhere to the follow- APICS ing disciplines and philosophies: Recognize that inventory and service level are the same KPI. Make inventory calculation a competency and part of the planning process. Capture and properly calculate data to support all system variables. Use these variables in a statistical model to determine what level of inventory is required to meet constraints. Use the model to determine how inventory might be scientifically reduced or right-sized. Proactively approach inventory targets with an improvement roadmap grounded in fact. Creating this type of model will enable much more proactive processes. This, in turn, makes it possible to communicate required inventory levels and understand how to reduce them. Most importantly, it helps business leaders avoid the ebb and flow so they may enjoy calmer waters. Steve Johanson is the chief executive officer of Supply Chain Toolworks and a founding partner of GTM Consulting. He may be contacted at email@example.com or (415) 533-9275. extra APICS Extra Live: Finding the Right safety stock and service levels Presented by: steve Johanson and lorine DeHuff Date: august 19, 2010 Time: 1:00–2:00 p.m. ct Attend APICS Extra Live to gain deeper insight into the July/August APICS magazine article by Steve Johanson, which illustrates how to achieve a balance between your safety stock and customer service levels. With good tools and processes, planners can simultaneously improve turns and service levels. In this APICS Extra Live, discover how effective planning considers sources of delay on both the supply and demand sides. Attendees also will learn why thorough safety stock calculation needs to be a part of the planning process and a fundamental competency. RegisteR online at apics.oRg/extRa. 46 July/August 2010 APICS magazine
safety stock targets. Most enterprise resources planning (ERP) systems perform a safety stock calculation. But very few include in the system all sources of variability as inputs to the safety stock formula. Furthermore, ERP tools rarely calculate accurate safety stock inputs or correct erroneous data. Figure 1 shows 13 basic safety stock inputs.
A VOYAGE CHAPTER L The Author’s reasons for undertaking a voyage of dis covery.—He builds a vessel for his purpose upon a new plan.—His departure from the United States. In the year 1817,1 projected a voyage of discovery, in the hope of finding a passage to a new and untried wo
Gulliver's Travels PART I p A VOYAGE TO LILLIPUT PART I PART II p A VOYAGE TO BROBDINGNAG PART II CHAPTER I. . Gulliver's Travels was published in 1726; and, although it was by no means intended for them, the book was . They cannot comprehend the occasion which provoked the book nor appreciate the
Least stable chair Most stable chair 1 1 1 Ð0.6 kcal/mol!G (c) For trans-1-chloro-4-methylcyclohexane (shown below), draw in bonds to CH3 and Cl groups as appropriate, to indicate the least and most stable chair conformers. Assuming that A values are additive (and hence subtractive if necessary.), predict what the ΔG value will be. (6 pt) Cl
Penn National Gaming Pinnacle Entertainment Penn National Boyd Gaming . MGM Resorts International Stand-Alone Stand-Alone Pro Forma for PENN/PNK merger Pro forma for recent acquisitions Consolidated and Pro Forma for acquisitions Consolidated IDR/IDCO b*/stable b*/stable --- B /Stable b */stable BB/Stable Statement Date 9/30/2017 9/30/2017 .
K. Webb ESE 499 15 Open-Loop Frequency Response & Stability Open-loop Bode plot can be used to assess stability But, we need to know if system is closed-loop stable for low gain or high gain Here, we'll assume open-loop-stable systems Closed-loop stable for low gain Open-loop Bode plot can tell us: Is a system closed-loop stable? If so, how stable?
The Tom Joyner Foundation Fantastic Voyage is The BIGGEST party with a purpose at sea.the Fantastic Voyage is a fully chartered, non-stop entertainment-filled cruise extravaganza! Sailing since 1999, the Tom Joyner Foundation Fantastic Voyage offers fantastic themed parties, activities, seminars, fitness classes, and top-notch entertainment.
Voyage is a 1 to 10-billion scale model of the Solar System that was permanently installed in Washington, DC, in October 2001. The real Solar System is exactly 10 billion times larger than the Voyage model. On this scale the Sun is about the size of a large grapefruit. The Earth is 15 met
Introduction to French Unit: Bon Voyage In the Bon Voyage Unit, students will explore travel in France through the use of authentic French language and culturally appropriate materials. During each lesson, students