【正文】 ocess is to develop an accurate, reliable view of market demand, which is called the demand plan. The Demand Planning process understands how products are anized and how they are sold. These structures are the foundation of the process and determine how forecast aggregation and disaggregation is conducted. A baseline statistical forecast is generated as a starting point. It is improved with information directly from large customers and channel partners through collaboration. The forecast is refined with the planned event schedule, so the demand plan is synchronized with internal and external activities. Each product is evaluated based on its lifecycle, and continually monitored to detect deviation. New product introductions are coordinated with older products, pipeline inventories, and ponent supply to maximize their effectiveness. Attach rates are used to determine ponent forecasts given the proliferation of products. The result is a demand plan that significantly reduces forecast error and calculates demand variability, both of which are used to determine the size of the response buffers. The specific response buffers and their placement are different based on the manufacturing model employed, therefore the Demand Planning process must represent those differences.The following figure identifies the key processes that constitute demand planning and the scenarios that are modeled in the template.Order PlanningDemand PlanningOrder PromisingAllocationPlanningDemand ForecastingTop down forecastingBottom up forecastingLife cycle planningOption forecast Consensus forecastingForecast extractionDemand CollaborationDemandPlanningCustomersOrder Creationamp。 i2。 ORB NETWORK。 CaptureForecast NettingMaster Planning7 / 36Demand ForecastingTopDown ForecastingDefinitionTop down forecasting is the process of taking an aggregate enterprise revenue target and converting this revenue target into a revenue forecast by sales unit/product line. This allocation process of revenue targets can be done using historical performance measures or using rule based allocation techniques. The revenue targets can further be broken down into unit volume forecasts by using Average Selling Price information for product lines.Historical information is typically more accurate at aggregate levels of customer/product hierarchies. Therefore, statistical forecasting techniques are typically applied at these aggregate levels. At levels where historical information might not be very relevant or is not perceived to be accurate, this allocation can be done with a rulebased approach. Frequency: This process is typically performed at a monthly/quarterly frequency, with the forecast being generated for the next several months/quarters. Scenario DescriptionBased upon historical bookings at an aggregate level across the entire pany (for all products and geography’s), the system will automatically generate multiple forecasts using different statistical techniques. The statistical techniques will account for such things as seasonality, trends, and quarterly spikes. Each statistical forecast will be pared with actuals to calculate a standard error. This will automatically occur at every branch (intersection) in the product and geographic hierarchies. The aggregate statistical forecast generated for the entire pany will be automatically disaggregated at every intersection using the statistical technique with the smallest standard error. The oute of this process will be a “Pickbest” statistically generated forecast at every level in the product and geography hierarchies. This forecast is then used as a baseline or starting point.Inputs ? Historical Bookings by units? Historical Statistically based Bookings ForecastOutputs? Multiple Statistical forecasts? Statistical “Pickbest” forecast? Forecast mitted to topdown forecast database row.Benefits? Easy disaggregation of data means faster, more accurate forecasting? Simple alignment of revenue targets? Uses top down statistical advantages to easily tie lower level forecasts to revenue targetsi2 Products Used TRADEMATRIX Demand Planner9 / 36BottomUp ForecastingDefinitionThis process enables the different sales anizations/sales reps/operations planners to enter the best estimate of the forecast for different products. This process consolidates the knowledge of sales representatives, local markets, and operational constraints into the forecasting process. This forecast can be aggregated from bottom up and pared to the targets established by the topdown forecasting process at the enterprise level. This will enable easy parison between sales forecasts and financial targets. Frequency: This is a weekly process. However, there is continuous refinement of the forecast at an interval determined by the forecasting cycle time and/or nature of the change required.Scenario DescriptionIn parallel with the topdown forecast, the sales force/operational planners will enter forecasts for independent demand for a particular SKU or product series by customer or region as is pertinent to a particular Product / Geography bination. This data will automatically be aggregated and pared to the targets established by the topdown forecasting process. Using the Average Selling Price for a unit, the unit based forecasts can be converted to revenue dollars and automatically aggregated.The bottomup forecast can also be generated using collaborative demand planning with a customer. In this case, the consensus forecast for a product/product series for a customer is aggregated and pared to the topdown target. Input ? Sales force input? Operations Planning Input ? Average Selling Price (ASP)? Customer forecast (from the Demand Collaboration process)Outputs ? Aggregated Sales forecast by unit?