The “Amazon Effect” has driven interest in Supply Chain efficiency, the ability to power inventory turns to new levels. Moreover, despite huge efforts forecast error has improved only slightly in the past 30 years. In fact, forecasts for consumer products are considered good if they are 60 percent accurate and very good if they are 80 percent accurate. Forecast error plus other process inadequacies such as Min/Max require huge amounts of Turns & Profit debilitating safety inventory not to mention the forecast error may consume your accumulated profit. The secret to efficiently powering turns was found in the US Army’s mandate for zero stock-outs and excess inventory. Clemson University developed a unique, simple, commonsense, time-based pull replenishment method utilizing the key principles of Theory of Constraints™, LEAN™, and Six Sigma™. Astonishingly, it mitigates forecast error and achieves unparalleled inventory breakthroughs in Turns, on-time delivery performance and inventory productivity while eliminating stock-outs and excesses.
The commonsense days-of-sales (referred to as days-of-combat) enable pull replenishment by factoring replenishment lead-times, stockouts, and buffer inventory simultaneously in terms of time. In other words, items in shortest time-based supply status (how long they are expected to meet demand compared with all other items) are replenished first to virtually eliminate stockouts with minimum safety inventory creating an efficient pull. While this method integrates key iTLS principles, it includes multiple other best practices and patented unique solutions. Users can count-on total elimination of stockouts, at least a doubling of inventory turns plus 30% reductions in inventory investment.
Background. The primary focus of inventory management improvements for over 30 years has been and continues to be improving forecast accuracy, but only a few improvements including demand driven MRP (DDRMP) and integrated Theory of Constraints™, Lean™, and Six Sigma™ (iTLS) have been made. While these clearly show great opportunities over standard systems, they have limitations to universal application. So today while dozens of sophisticated forecast-based solutions exist, consumer product forecasts are still considered “good” if they are 60 percent accurate and “very good” if they are 80 percent accurate.
Obviously, forecasts are needed so products can be ready when consumers want to buy them and more so now with the introduction of the “Amazon Effect.” A major reason for inaccuracy is that consumer needs change, but there are other major unsolved inaccuracy drivers.
The inventory management process begins with the generation of a high-level rolled up sales forecast as the basis for annual budgeting and capacity planning. Forecast accuracy is generally acceptable at this level if the propensity for padding at multiple management levels is controlled properly. However, not so when the forecast is exploded to the individual SKU level for replenishment (what to order, make, and move). Forecast error and resulting problems are greatest the closer to individual SKU level and over the shortest time horizons; the environment within which inventory managers must make replenishment decisions.
Root Cause Problems. Natural variations in demand and replenishment lead-times as well as the impact of unmanaged resource constraints, all of which drive forecast error and unacceptable stock out rates, are highest in this environment. At least 16 root cause variation drivers are present in all manufacturing and replenishment operations. So, you see, it’s not the manager’s fault, standard systems and practices are not capable of solving these problems, so managers use the universally accepted solution of adding more and more unaffordable safety inventory to all SKUs to reduce fewer and fewer stockouts. Simply stated, forecast accuracy has reached its plateau with standard systems and practices so a different method must be used to optimize inventory performance.
The Solution. Replenishment pull is widely recognized to be superior to forecasting and pushing primarily because it eliminates forecast error. What consumers buy today is the best forecast of what they will buy tomorrow. A well-known example is Burger King’s “Have it your way” strategy that doesn’t create excess thereby reducing costs and increasing quality in exchange for a short wait. However, pull is seldom used because consumers are not willing to wait the required time and standard systems and practices lack the concepts and tools required to eliminate the long wait times. In fact, we found all known best practices and some unique solutions had to be integrated to enable pull to work by minimizing replenishment lead times and large buffer inventories for all SKUs.
The Pull Signal. Consumers initiate demand and are the only supply chain participants who send a pull signal each time they purchase an item. Everyone else up the supply chain accumulates customer orders to pass them upstream in larger and larger batches infrequently as their replenishment orders. Thus, the daily selling of an item is the ideal replenishment pull signal for order and make entities as the longer the replacement cycle, the larger the point of sale inventory must be.
The Pull Response. inventory must be on-hand to ship immediately upon receipt of the pull signal, but not in the huge quantities for all products used today. This takes us from the symptoms; e.g. stockouts and excess inventory through root causes; e.g. replenishment lead time and demand variations to the core problem.
The Core Problem. The core problem is determining which SKU in what quantity should be replenished next to minimize both stock out and overstock risks while honoring resource constraints. Clearly, the next SKU replenished should always be the one in shortest supply status simply because there is no way to know which one will need replenishment next. So, how do we determine this?
The Primary Solution - Days of Sales. Selling products not only provides the ideal pull signal, but also the key core problem solution. Standard systems replenish SKUs individually and independently of all other SKUs because they contain no relationships between SKUs. Since time is the natural common denominator of all activity, we use it from the forecast to create a projected day-of sales demand quantity for each SKU. This is forward, not rearward, looking and sufficiently accurate to serve as the relative demand common denominator between all SKUs. With this the SKU in shortest days-of-sales status can be easily identified and replenished first. But, what should the upper replenishment limit be for each SKU?
Replenishment Objective. Each SKU must be replenished to a Days-of-Sales objective which consists of:
1. A small on-hand safety inventory to enable immediate shipment, level replenishment ordering, and absorption of demand and replenishment lead-time variations. This normally begins with an RLT of coverage adjusted as necessary based on status to minimize stock out risks upon implementation.
2. Replenishment lead-time days of coverage sufficient for incoming orders or work-in-process.
If there are no constraints, enough inventory is simply added to all SKUs to reach the Days-of-Sales objective. However, constraints always exist and should always be managed to avoid unnecessary costs. For example, cash, credit, time, capacity or other constraints often preclude replenishing to the Days-of-Sales objective so we must avoid over replenishing some SKUs and under replenishing others.
Constraints Management. Scarce resources must be allocated first to the SKU in shortest days-of-sales supply status in a manner to minimize and equalize stockout and overstock risks for all SKUs. To achieve this, we allocate the lowest permitted batch quantity to the SKU in shortest days-of-sales supply status and repeat this process until the objective is reached or the constraint capacity is exhausted. This leaves the on-hand plus due-in or work-in-process orders balanced in days-of-sales for all SKUs.
Balanced SKUs. Simply achieving low level SKU balance enables the virtual elimination of stock outs and about 30 percent reduction in unnecessary inventory. Once all SKUs approach balance in Days-of-Sales, other opportunities for additional improvements become feasible.
SKU Velocity. The next major opportunity is because high demand SKUs are easier to keep in stock than lower demand SKUs because they are ordered, made, and moved every time any SKUs are replenished. Once on-hand plus due-in inventories are balanced, outgoing replenishment orders are automatically balanced, and the balance is normally well maintained for incoming replenishment products. While physical replenishment lead-times might be 30 days, weekly balanced product shipments that match normal weekly manufacturing cycles reduce the effective replenishment lead-times to 7 days for high demand SKUs which are often 50 to 80 percent of all demand. This enables the effective or “virtual” RLT to be much less than the physical RLT for most of the SKUs.
Supporting Solutions. In total, 12 integrated solutions are used to achieve the above and following improvements:
· Obtain current inventory status visibility in days-of-sales.
· Minimize demand and replenishment lead time variations, and over committing of constrained resources.
· Minimize the risks of stockouts and overstocks at the same time.
· Frequently and automatically calculate optimum, actionable replenishment quantities.
· Minimize forecast errors by always being in stock to capture all demand.
Together these solutions provide the commonsense logic with the simplicity to readily focus all managers on the right objectives for achieving the critical goal of maximizing customer service while minimizing inventory investments and expenses.
An Example. Here’s a scenario for applying the BI replenishment methodology to a sales plan. A distributor’s replenishment lead time is 90 days for a product family of widgets that collectively has forecasted demand of 1,000 items per month.
Minimizing replenishment lead times. Since computing replenishment computations in days-of-sales enables balancing all SKUs to the same days-of-sales status, a one-third inventory reduction is readily achievable by eliminating excess inventory. We first set the beginning on-hand inventory target at 60 days-of-sales (one-third below standard replenishment stockage targets) and the replenishment lead time at another 60 days-of-sales for a total objective of 120 days-of-sales on-hand plus on-order. Then, after a few replenishment cycles, we will see a flow of balanced orders going out and balanced product shipments coming in. This balance enables even further stockage reductions of up to a second 30 days-of-sales and sets the stage for additional improvements.
Synchronizing Replenishment to Sales. As these replenishment lead time days and inventories are being reduced, days-of-sales computations frequently check SKU sales velocities making automatic adjustments for the proper lead time into the future. This minimizes both stockout and overstock risks at the same time.
Complex Solution – Simple Implementation. The only known method of making a dramatic breakthrough in inventory performance requires shifting from forecasting and pushing to pulling inventory based on demand. Making pull work requires the integration of all known and several new focused and unique inventory optimization solutions. Most of these solutions are integrated into a program using standard MRP or MIN/MAX input data so that the complexity is handled by the program to compute vastly improved replenishment requirements. Users simply replace the outputs of their standard replenishment computations with these new requirements in a manner that can be implemented in a few hours and run in a few minutes each day.
Proven Performance. Balanced Inventory is an advanced integrated Theory of Constraints, Lean Manufacturing, and Six Sigma (iTLS) based tool developed for the U.S. military and proven by the U.S. Army to perfectly synchronize combat critical products to warfighter needs. In doing this it met three specific objectives that could not be approached by standard methods:
1. Field a new combat system from specialized raw material production through multiple end-item production 25 percent of shortest ever previous time.
2. Have zero stock outs of raw materials, components, and end-items.
3. Accomplish this without building up the massive inventories required by standard systems and practices.
This is a SaaS program developed at Clemson University’s advanced manufacturing facility for the military and is now commercially available and designed to produce on-going benefits of at least ten times it’s annual fee. Typical results include stock-out and associated expense elimination, over 30% inventory reduction with at least a doubling of turns, and at least a 30% reduction in inventory expense and management costs.
Implementation Steps. Balanced Inventory was crafted as a tool to synchronize supply chain-wide order, make, and move replenishment actions with consumer demand. It begins with local implementation to demonstrate clearly the existence of improvement possibilities never imagined. It is structured for easy extension to vendors and customers, thereby improving the profitability of all supply chain partners as they work together for their mutual benefit.
How We Can Help?
Has complex global sourcing coupled with multi-channel sales have you conceding inventory turns and profitability? Are you constantly expediting to meet completions or stay in-stock? Are you filling containers with product you really don’t need? Having trouble finding Inventory Analyst that understand your business? Are you in spread-sheet Chaos? Is your CFO concerned about tight cash?
Implementation begins with an analysis of potential improvements with the first balancing run of a family of products carefully chosen to represent the “center of the business” and moves to a full run of several product families representing the “breadth of the business.”