If your firm is like most, it is much easier to add a new SKU (stock-keeping unit) into the product line than to drop one. In fact, in the large majority of manufacturing and distribution firms we have worked with in the past, it seems to take the proverbial "act of Congress" to eliminate underperforming items. And, while well-intentioned managers will agree in concept that the overall product mix should be monitored and periodically purged of the "cats and dogs," when a list of specific SKUs is developed to meet the needs of one business function or another, the battle lines are drawn. Typically, sales and marketing managers want to retain as many SKUs as possible, while operations managers want to keep as few as possible. Unfortunately, most firms do not have an objective process for ongoing SKU rationalization. Frequently, inventory investment and associated carrying costs are some of the measures used to evaluate SKU performance, along with other factors, many of which are subjective. However, beyond just the costs associated with inventories of underperforming SKUs, there are many hidden costs that should be considered in evaluating the true economic value in retaining the item as an actively supported product. These hidden costs include expenses associated with maintaining data files, physical and/or online catalogs, engineering drawings, revision levels, purchasing records, cycle counts and so on, which should be (but typically are not) considered in the calculation of carrying costs. A consistently applied, objective process for evaluating SKU performance and periodically rationalizing the product line can help to ensure that the visible costs as well as the hidden costs associated with actively retaining items in your product line are considered. This column will discuss a technique that Cap Gemini Ernst & Young has used successfully in the past with fill-from-stock manufacturing and distribution clients in managing the trade-off between inventory investment and fill rates, as well as identifying potential candidates for eliminating from the product line. I have named this technique the two-dimension inventory efficiency model. Two-Dimension Inventory-Efficiency Model Analysis At its core, the two-dimension inventory efficiency technique is an analysis of usage value and order frequency applied to the product mix to determine current inventory investment vs. fill rate performance. The analysis can be applied to the entire product line and/or sub-segments, such as by distribution center, by product family, by sales channel, etc. The result is an indication of the relative inventory investment, turns and fill rate for each of nine usage value-order frequency segments of the products in question. The 2-D inventory efficiency model can be best visualized as a 3-by-3 matrix, in which there are nine segments. One axis of the matrix represents traditional APICS-style A-B-C categories of annual usage value (AUV), in which the AUV for each SKU is calculated as the annual demand multiplied by the cost. Then the SKUs are sorted in descending AUV, with those SKUs comprising 80% of the cumulative AUV being considered "A" items, those SKUs comprising the next 15% of the cumulative AUV being considered "B" items and remaining SKUs being considered "C" items. For the purposes of this type of analysis, AUV is used as a proxy for revenues and to minimize the impact of variations in gross margin between SKUs or groups of SKUs. The second axis of the matrix represents a categorization of SKUs by order frequency. That is, those SKUs that are ordered by customers most frequently are in the "HI" category, those SKUs that are ordered least frequently are in the "LO" category and those SKUs in between are in the "MED" category. While it would be ideal to have precise definitions for "HI," "MED" and "LO," the practical application of this technique dictates that each company's situation calls for different delineations. For one client (a distribution company), SKUs that fell into the "HI" order frequency category were those that were ordered by customers at the rate of more than 63 times per month. For another client (a specialty chemicals manufacturing and distribution firm), SKUs that fell into the "HI" order-frequency category were those that were ordered by customers at the rate of more than 21 times per month. In general, application of this technique seems to work best when approximately 30% to 35% of SKUs fall into each category, with not all categories necessarily having to be equal in this regard. Interpreting The Results When we segment the total population of SKUs into the nine segments according to the two dimensions described, typically some interesting observations can be made about the composition of the inventories associated with each group of SKUs. For example:
- Unless inventories are significantly distorted, we would expect that the inventory turns in the upper left four segments of the matrix ("A" & "B", "HI" & "MED") would be greater than those in the remaining five "non-core" segments. The SKUs in these four segments can be thought of as the "core" SKUs of the business because these SKUs contribute the significant majority of revenues (typically 80% to 90%) and the majority of customer order activity (typically 70% to 80%). For the specialty chemicals client mentioned earlier, of 12,400 stocked SKUs, the 10% of SKUs in the core segments provided 89% of AUV and 75% of the customer order activity.
- The SKUs in the core segments provide the most effective short-term opportunity to improve inventory efficiency, either by reducing inventory investment required to support target fill rates or by increasing fill rates without increasing inventory levels.
- Core SKUs can be managed for both high inventory turns and high fill rates as they are in constant demand and the demand levels generally tend to have lower variability than non-core SKUs.
- Non-core SKUs generally will have poorer inventory turns rates than core SKUs and, in some cases, can and should be managed with less sophistication or organizational energy, as these SKUs contribute relatively less to revenues and overall customer order activity.
