The Key to Understanding the Forecasting Process

Information is a company's key asset ...an interactive process with the different business functions is the proper mechanism for achieving accurate forecasts...Senior Management will give full support to the forecasting function if they understand its value. Forecasting, for most companies, is the central process from which Finance, Marketing, Sales, Logistics and Senior Management direct and coordinate their efforts. Over the past 15 years, I have had the opportunity to review and participate in the forecasting process and forecasting activities within various Fortune 500 companies. What I have discovered is that forecasting within most companies is a disaggregated function that results in key personnel directing themselves in an uncoordinated effort to achieve the company's sales and operating plan. WHY FORECAST SALES Recently, on a flight from New York to Chicago, I found myself sitting next to the Chairman of a large Fortune 500 consumer goods manufacturer. I questioned him regarding his company's greatest assets. One key asset he stressed was information. In asking him why information is of such great value, he said, I understand what has happened and why it has happened, I can then successfully predict what will happen. If I know where I am going and can affect that, my company will enjoy a significant competitive advantage. In discussing his company's effectiveness to create information from their huge data investment and realize the desired competitive advantage, he admitted their inability to accomplish this consistently. experience conveyed to me by this Fortune 500 Executive is not different from what I have discovered to be typical of most Fortune 500 companies. question I continue to ponder is, Why have the vast majority of forecasting systems implemented within various corporations over the last several years fallen short of providing the value and return on investment that was expected of them. When I discuss forecasting with forecasting professionals, invariably the conversation focuses on the accuracy and capabilities of the forecasting techniques that they employ. They seem almost obsessed with what I have come to call The Magic Model. Forecasting projects a picture in most peoples' minds similar to that illustrated in Figure 1. (Figure 1 omitted) Figure 1 graphically displays actual demand, predicted demand (based on the utilized or chosen forecasting model) and residual values, which quantify the difference between actual and predicted demand. end product of the graph is a forecast which is generated based on the model utilized to calculate predicted demand (the historical forecast). I have seen dozens of forecasting models and systems that do an excellent job of predicting historical demand and create a future forecast accordingly. However, even with the variety of different forecasting techniques, and various computer technologies available today, for the most part, forecast error remains significant, and effective forecasting within major corporations is seldom realized. result of these experiences raises questions regarding the real importance of the forecasting technique(s) and its contribution to establishing an effective sales forecasting system. Most forecasting professionals agree that the goal of forecasting is to consistently predict future sales. This results in improved planning and more efficient utilization of corporate resources, including raw materials, production scheduling, logistics planning, inventory management, capital expenditures, promotion spending, and the overall budget requirements necessary to support the forecast. To accomplish this, most companies set out to develop a forecasting system consisting of three major components: a forecast algorithm(s), forecast data/database, and an easy to use system interface for generating the forecast. As depicted in Figure 2, to create a forecast, the end-user, generally through some graphical users interface, retrieves the required data from a forecasting database, streams the data through some forecasting algorithm(s) and then reports the data in both tabular as well as graphic format. …