Brian Dreckshage04.12.07
Continuous Replenishment Helps Distributors Reach Inventory Balance
By Brian Dreckshage
When it comes to having enough inventory to satisfy customer demand, many distributors struggle to strike a balance between having too little or too much. Most operating systems can help manage inventory but are not often configured to optimize it; however, continuous replenishment (CR) is an optimization process that is.
CR Applies Lean Principles
The goal of CR is to bring in inventory when the customer needs it. The process applies Lean principles to distribution operations as a way to increase inventory velocity. As opposed to the old “receipt, store and pull,” CR can help your warehouse function closer to a cross-docking operation where items are received, then staged to go out.
The systems you commonly use can serve as a base for implementation, but CR mainly relies on good management processes to succeed. Though you may not be able to convert all your items to CR, you could benefit by converting just your top sellers.
Benefits of CR include increased cash flow, reduced warehouse labor costs and the possibility of lower warehouse space requirements and costs. As mentioned, a successful CR process combines system configurations and effective business processes. The following methods and strategies support CR and help it succeed.
Forecasting
Most distributors sell goods from stock and must stock their inventory unless they can order goods from vendors and receive them in less time than a customer’s expected lead time.
Stocking inventory requires forecasting, but most systems lack forecasting tools; thus, stocking often is done without them. However, with personal oversight by sales management, forecasts can be produced using standalone statistical forecasting tools.
Forecasts can help shape your demand plan, but don’t confuse forecasting with demand planning (see below). All forecasting should do is predict sales, which are only one part of overall demand.
Commonly used statistical forecasting tools offer “best fit” algorithms and accuracy measures, as well as deviation statistics used in establishing safety stock levels (covered later in this article). If you currently have a system tool for forecasting, it’s important to understand how it works.
As stated, best-fit forecasting software usually can create a forecast that matches product demand patterns. This means a one-size-fits-all approach to forecast modeling not only is inappropriate, it is potentially disastrous in many cases.
Your forecasts, as such, always will be off to some degree, but the real key to successful distribution is managing inventory so you have enough product to meet customer expectations without having too much. This balancing act isn’t easy, but it’s your primary value-add in the supply chain. Reasonable forecasts, along with variation-based safety stock, will help you achieve balance.
Demand-Planning Strategies
Distributors often don’t spend enough time demand planning, and their customers can suffer for it. Simply put, they must get beyond treating demand planning as something to pawn off on junior staff members. At the very least, it should rank in importance with customer service, especially because the quality of service depends on how well demand planning is done.
Demand planning is the culmination of every item you need to service your customers versus what’s actually available to sell. In addition, demand planning can expand to include:
• Inventory stocking plans that typically represent an inventory build-up to satisfy seasonal peaks
• Account planning, which considers factors a forecast algorithm does not know, including customer promotions, customer inventory build-up/draw-downs, etc.
• Capacity or how your customer delivery may be compromised by vendors or your own operations
• Customer service levels; a 100% fill rate may be unrealistic, so determine the level you can or should reach
Additional Planning Methods
Reorder point (ROP) and materials requirement planning (MRP) are methods that can work successfully on their own or in combination with most systems.
You typically use ROP when you don’t want to forecast demand directly; however, you do need to know—by item—your “average” sales (demand) and lead times, which often include service parts or mid- to low-volume items with low value.
With ROP, you are able to monitor your available inventory—ie, what you have on hand minus allocated inventory. When this drops below a certain point—either the reorder point or minimum (MIN)—a planned order is triggered.
For example, most software won’t calculate ROP for you, but the easiest way to configure it is with this formula:
D (demand during lead time)
+
S (safety stock)
If you sell an average of six items per week for an item with a three-week lead time, D will equal 18 (safety-stock details follow).
Although MRP works in a similar way, this method is forecast driven; you start with on-hand inventory and subtract forecasted demand to arrive at a projected available inventory number, “projected” because you are looking one or several forecast periods into the future, depending on your product lead time.
Over time, actual sales numbers are consumed from the forecast (projected available), and planned/open purchase order (PO) quantities and PO receipts are added to the projected available. When the projected available falls below the MIN (safety stock), a new planned order (to buy more) is generated.
The key to CR is in the delivery dates for planned orders, specifically with MRP items. Most MRP systems strive to have forecasted quantities on hand at the beginning of the month. If your MIN order quantities allow it, manually spread these delivery dates throughout the month. Some systems will do this if forecasts can be converted to weekly buckets, but most systems do not.
Also, MRP usually will generate actions—eg, move POs in/out, increase/ decrease order size or cancel unnecessary orders. Make sure someone responds to these action items.
You can determine your safety stock in a number of ways, the most common method being 50% of D, which often leaves distributors grossly overstocked. A preferable method or model calculates safety stock using the standard deviation of demand, which normally is available as a statistic from your forecasting program.
Other strategies include:
• Reduction of order quantities. Work with vendors and make more frequent MIN orders, which helps turn inventory more quickly. This option may not apply to A items but will be more appropriate for B and C (slower moving) items.
• Reduction of lead times. Long lead times won’t necessarily hamper CR, but they can limit flexibility when demand changes, as it unrelentingly will. Work with your vendors to reduce lead time as much as possible. Using a purchase forecast (via blanket order) with vendors can assure you of capacity and allows you to vary the mix up to the vendor’s freeze fence.
• Adjusted safety stock levels. If you experience seasonal or cyclical demand, adjust your safety stock levels to match the major ups and downs of demand. This provides more safety stock during heavy demand periods and cuts back inventory during slowdowns, which helps increase inventory turns and free cash flow.
• Demand plans and forecasts. Because theses are subject to change, be sure to update these major system inputs as warranted. Remember, the key to success in CR is your ability to react quickly to change.
• Inventory turns. If not already a key performance indicator, make sure you include them. Measure inventory turns not on sales but on a cost-of-goods-sold basis and prioritize by ABC and inventory dollars. Low turns usually pinpoint planning or purchasing inefficiencies.
* * *
Continuous replenishment can help you increase inventory velocity both in and out of your facility. While it’s difficult to apply “pure” Lean principles to a distribution environment, the ability to set up a “built-for-speed” CR model has proved profitable.