Use of Volume-Variance Demand Analysis in Planning Strategy Formulation
Alan L. Milliken · 2010
Shows how to combine volume analysis with variance analysis to determine the contribution of a product to the company 's profit ... product with a high variance because of seasonality and data trend is statistically predictable ... collaboration with a customer helps to improve forecasts and reduce inventory. The downturn in the global economy has encouraged firms to take a closer look at their excess, slowmoving, and obsolete inventories and assess why these exist. Cash is King in this economic environment. Management's bias toward lower inventory, even to the point of sacrificing service and cost, has resulted in much more pressure on supply chain to find and correct the causes of dysfunctional inventories. One means of better identifying the inventory risks associated with offering products on a make-to-stock, off-the-shelf basis is vo lume- variance demand analysis. WHAT IS VOLUMEVARIANCE DEMAND ANALYSIS? Of course Pareto 's Law has been used to identify the significant few for centuries. Supply chain has performed ABC volume analysis for decades. For example, ABC segmentation has been used to assign stocking policies or to determine cycle count frequency. More recently, firms have begun to use ABC segmentations to evaluate customers from the perspective of total cost-to-serve and products on the total cost-to-acquire basis. The latest use of ABC volume analysis for products is in combination with variance analysis. Supply chain also has many uses for variance analyses. For example, standard deviation is a key input to statistical safety stock calculations. In the past, the evaluation of demand variation was limited to the use of sophisticated statistical techniques like R Squared or Correlation of Coefficients to determine how well a forecast related to the historical data. Now, more and more firms are using the Coefficient of Variation (CV) to evaluate the probability of successfully forecasting a historical demand pattern. USING THE COEFFICIENT OF VARIATION In probability theory and statistics, the coefficient of variation (CV) is a normalized measure of dispersion of a probability distribution. It is defined as the ratio of the standard deviation to the mean. Distributions with CV>1.0 are considered high variance while those with CV It is not a great leap of faith to accept that in most cases if the variance is high, statistical forecasting tools designed for normal, bell-curve distribution may not perform well. This is particularly true for averaging and smoothing models, which are the most used in practice. Therefore, one use of CV analysis is to determine the probability of forecasting a historical dataset within a reasonable limit. Again, most firms accept the threshold of C V> 1 .0 in segmenting those items that cannot be forecasted with statistical forecasting software. In the dataset given in Table 1, two products did not pass the CV COMBINING VOLUME & VARIANCE ANALYSES Many firms combine volume analysis (ABC) with variance analysis (XY) to determine the contribution of the product to the value of the firm (revenue, profit) and the probability. Their forecasting software can provide a reasonable forecast to control inventory and improve cash flow. The results of an ABC-XY analysis are usually displayed in a matrix chart as well as a table of values. …