It is important to note that the above applies if inside the process is a structured process of forecast. Sales and Marketing Predemanda many processes for demand planning meetings follow the following process: generate statistical forecasts. Adjust the forecasts based on the knowledge of the market. Reaching consensus and publish the results. A better strategy would be to bring the knowledge as an input in the development of statistical forecasts. This strategy has had a great success among many academics forecast. J. Scott Armstrong (2001, p.
736), in his book Principies of Forecasting, mentions the importance of using knowledge of marketing and sales as inputs in the development of a functional planning. Time Warner will undoubtedly add to your understanding. Principle 11.2: Use of forms structured knowledge as inputs to quantitative models. 11.3 Principle: Use the knowledge to select, assess, and modify quantitative methods. The investment of time and energy in models that incorporate knowledge properly will help analysts to create more accurate statistical forecasts, It will reduce the number of manual modifications, and improves the accuracy of forecast. At the J.R.
Simplot company, sales and Marketing staff participating in meetings of predemanda. At these meetings, they organize, evaluate, and formula the last commercial information. It documents the process fully and shared the results with analysts before the construction of statistical models. Let’s see some resulting conclusions in this process of incorporating knowledge of sales and Marketing who helped improve the statistical prognosis results: to the forecasting statistically level SKU and then distribute the results in descending achieved better results than pronosticando of form desagragada. Carry out a classification of pareto on references allows to concentrate our efforts in the references A type (those critical business) and schedule revisions with frequencies lower than in the type B and C. Articles whose behavior is too variable for forecasting through a model statistical, are removed gradually as the months pass since usually these articles present special features such as contracts, amounts and fixed dates, new items which do not have history, which makes the performance of a statistical model very poor.
{ Comments are closed! }