Forecast Generator

Estimating customer demand is key to any business so that orders can efficiently bet met. Forecasts play a key role to start the demand planning process, however, most statistical forecast are often wrong.  The process of incorporating qualitative business intelligence into actual planned forecasts is crucial in making forecasts more accurate.

planLM Forecast Generator generates forecasted orders by determining the best fitting statistical function based on the order history and specified external factors, and extrapolating this function into the future.

Features

Statistical Methods

planLM Forecast Generator runs a full range of well studied statistical methods, including multi-linear regression, single/double/triple exponential smoothing, multiplicative Winters, Box-Jenkins, and Croston methods.

Most of these statistical methods require parameters and these parameters effect the accuracy of fit. planLM Forecast Generator has the capability to auto-adjust these parameters for best results.

planLM Forecast Generator also features "Pick Best" option to automatically select the best statistical method among all available methods.

Aggregated Forecast

Generating statistical forecasts at higher levels of aggregation is often more accurate in aggregate. The aggregation levels for product, sales channel, customer and time dimensions are fully customizable in planLM Forecast Generator. Once the aggregated statistical forecast is generated, it is converted to the lowest granularity by utilizing proprietary disaggregation algorithms.

Multi Scenario Support

planLM Forecast Generator works in conjunction with "What If" capability of planLM data model, as well as the multiple forecast version can reside within a scenario.

Scalability of 64bit architecture

Fully utilizing 64bit architecture, there is no operating system based practical limitation on the size of the statistical model.