Accurate inventory management takes more than a good software package.
The most expensive, sophisticated software package will not automatically result in an optimal level of inventory for an electrical wholesaler. “Optimal” means a high level of customer service and inventory turns, but with low inventory investment. To achieve and maintain an optimal level, employees educated in the principles of effective inventory management must understand how to set certain parameters; then set them and keep them set right. Here are some tips about principles of inventory management and setting key parameters.
Too often I work with users who have been trained in the rote mechanics of using the ERP software package that literally runs the distributorship, but were not educated in modern inventory management principles. Important principles include: the relationship between customer service level and inventory level; the meaning of the normal statistical distribution (aka “bell curve”) that plays a role in the calculation of safety stock and the adjustment of data on odd sales events (e.g., a very large quantity of wire sold from stock). A principle related to purchasing, and so indirectly related to inventory management, is the line point (LP). This is not another term for order point (OP). The LP is the OP plus sales forecasted for the buying cycle (time between buys). Items above their OP but below their LP should be purchased only when those items are needed to make a purchase minimum, or would result in a purchase discount or freight allowance that would be larger than the cost of inventorying those items for longer than usual.
Another basic that is sometimes skipped is the re-setting of parameters. When the system was first installed, users were too busy to determine what values to set parameters to; so the system went live with default values (that are on average good for all wholesalers, but not good for any specific electrical wholesaler). Of course, users are still too busy to investigate the values and change those that are not right for the company if they could even do so without first learning the principles of effective and profitable inventory management.
Qualifying historical data
Although most systems adjust historical data to remove oddities before using the data to forecast future sales, the scope and amount of an adjustment depends on the values of certain parameters. In addition to the common oddities of sales spikes (large, one-time sale of EMT from stock for a job) and dips (a customer returns large quantities because a job is stopped), there can be periods of no sales. Sales spikes can also be caused by wholesaler promotions, perhaps followed by decreased sales because customers bought ahead — two kinds of oddities. The values of parameters determine whether an oddity will be adjusted, and the extent of that adjustment (which may increase with the size of the oddity). Users should not re-set these parameter values until they know what all the oddities have been and are likely to be; which parameters adjust data; and how the value of each parameter changes the adjustment.
Many systems come with several different formulas for forecasting future sales by using history. One of those formulas is the default — the one that will be used unless someone selects another formula. Life would be easy if one formula, the default or otherwise, could be used for all items, but that is very rarely appropriate. Even the use of one formula for all items in a particular product-line would save time, but that too is seldom appropriate, because every product-line has slow-moving items such as a 3-inch pipe coupling, and they cannot accurately be forecast with the same formula that works well for fast movers such as a 1-inch pipe coupling. As with the re-setting of most parameters, it is necessary to select different formulas for different items, unless the system can automatically select the best formula (based, of course, on parameters that define best). Formulas that are easy to understand but not accurate include averaging and weighted averaging (where users set the weights, the emphasis factors). Wherever possible, use the more sophisticated formulas, even though someone still needs to re-set related parameters.
If a system measures the accuracy of an item's forecast (sometimes called the mean average deviation, or MAD), accuracy reports should be reviewed quarterly to determine if parameters should be changed or a different formula used.
EOQ is dead, long live EOQ
For some items, EOQ (Economical Order Quantity) is inaccurate: items with a very low unit value relative to the cost of procurement (wire connectors); or items that sell infrequently (120V fuses). For these kinds of items, EOQ would calculate a multi-year supply or a quantity of zero, respectively. A better way to handle both kinds of EOQ-inappropriate items is to use the dynamic min/max feature of the system, whereby the system uses history to determine the values of min and max. But before doing so, research the system's min/max formula, and determine what the min/max parameter values should be, and if min/max would produce realistic results. Minimize the use of manual min/max because it's not dynamic, and so takes a lot of effort to keep current as sales patterns change.
Safety stock too often accounts for a large portion of an item's quantity on hand. And for too many items, the quantity on hand is never less than the level of safety stock — which means that the safety stock is never used and is dead inventory. One reason that related parameters are sometimes set wrong is that some people do not understand principles for calculating safety stock: 1) safety stock is kept in case sales exceed forecasted sales; 2) the level of safety stock does not depend on an items velocity; 3) the level of safety stock for an item should be mainly in proportion to the volatility of its activities; 4) the level of safety stock should be based on the item's target service level.
Lead time may be the most difficult value to determine, because it is basically beyond a wholesaler's control, and because some lead times are seasonal, even though sales of the items are not (some factories still shut down for a month during the summer). But that is no excuse for not examining the default values of related parameters, which are often set with the assumption of constant lead time. Where a system contains optional sophisticated formulas for calculating lead times, those formulas should be investigated, compared to vendor performance, and used wherever possible. Even if there are no sophisticated formulas, related parameters should still be set in the context of expected future vendor performance.
Dick Friedman is a recognized expert on warehouse operations, management and technology for electrical wholesalers, but he does not sell computer systems or warehouse technology or equipment. Based on more than 30 years of warehouse experience, he helps wholesalers reduce warehouse errors and increase productivity; often through inexpensive changes to operations, management or technology. He is a contributing author to Electrical Wholesaling, and consults with readers. Call (847)256-3260 for a free consultation about improving warehouse accuracy and productivity or visit www.GenBusCon.com for more information or to send e-mail.