Friday, September 2, 2011


Forecasting is a technique for finding the future needs, by examining the historical data and trend. It is an unbiased estimation for future demand of various variables such as sales, production etc., on the basis of past data and experience. Simply, extending the past data into future using time line is not forecasting, judgment is very much essential for forecasting (Kurtz 2010). Forecasting technique is imperative to improve the communication in the organization as forecasting requires information from the different departments. Cross functional approach of forecasting is used, where independent forecasting group conduct the forecasting by integrating different departments (Mentzer and Moon2005). This communication reduces the duplicity in the forecasting efforts and used all relevant information.

Sometimes, forecasting becomes meaningful and the actual output is not the same as expected. There are two basic reasons behind this viz. under estimation and over estimation. In under estimation, it is not possible to achieve the forecasted results due to insufficient resources and dissatisfied customers, while in over estimation the forecasting fails due to the under utilization of resources or spending the resources too early (Kurtz 2010).

To make the forecasting effective, organizations have to effectively allocate the resources and take proper measures that too in regular time interval to equalize the actual results with the forecasted outcomes. The forecasting must be done by analyzing the data properly and after finding the result, it must be evaluated back to find the feasibility of the results. Also, the forecasting must be made effective by proper checking the actual work based the forecasted results (Halder 2010).

Forecasting with the help of Auto Regressive Integrated Moving Average (ARIMA) and autoregressive (AR) models typically produce better results than those of simple models as it consider the problem more precisely and produce results that can be closer to the actual results. These models sub-divide the give data trend into various types such as constant trend, linear trend or quadratic trend that helps the researcher to accurately forecast the outcomes (Maindonald and Braun 2010).

For reading - Forecasting Steps

Blog Post Provided By -
Contact Us Email Id -