Good Example Of Sales Forecasting And Budgeting Essay

In this essay we will perform different approaches to make a forecast of sales volume in 2010 year with the given data. Modern statistical forecasting methods are exponential smoothing models, with an autoregressive moving average of econometric equations based on both parametric and non-parametric approaches on.
At first glance, to decide which method is appropriate to use for this data it is better to look at the line graph:
It seems that 3 year weighted average method is appropriate to use here. However, I check all methods mentioned in the conditions of the task.
Moving average method is easy enough to use, but it is too easy to create an accurate forecast. Using this method, any forecast period is not nothing like getting average results of several observational time series. For example, if you select a moving average over three years, the forecast for 2010 is the average value of indicators for 2007, 2008 and 2009. Taking as a method of forecasting the moving average for five months, you will appreciate the index in 2010 as the average figures for 2005-2009 years.
For simple average I just expect that the prediction value is equal to sample mean value. I use Excel and calculated that this value is equal to 1765:
900+1400+1750+1900+1700+1500+2000+1950+2250+230010=1765

For three year weighted average, the formula to calculate prediction is:

Ft=3At-1+2At-2+At-36

Where F is a prediction value and A are the values of the data set.

According to this formula, I’ve got that the prediction value of sales in 2010 is 2225
For five-year weighted average I use the following formula:
Ft=5At-1+4At-2+3At-3+2At-4+At-55+4+3+2+1

According to this formula, the volume of sales in 2010 is approximately 2123.33

One of the most effective ways to short-term forecasting of demand forecasting is based on exponential smoothing. This method is based on the calculation of the weighted average sales for a number of past periods and in the short term is very useful in cases where the time series in the present there are significant differences in the data, as well as in cases where the goods cheap or perishable.
The advantage of the method of exponential smoothing with short-term forecasting consists mainly in the fact that it is quite simple and easy to use compared to other methods. In addition, the exponential smoothing forecast provides rapid response to all events that take place over a period that allows you to build the so-called "Adaptive predictive model." This model allows considering random variability of the function much better than a trend model. Nevertheless, demand forecasting exponential smoothing method will be more accurate if the determinants of demand, not subject to sharp fluctuations.

For a five year exponentially smoothed average, the formula is:

Where c is a value from given set, s – are the values of the forecasting data set. And parameter “a” is a value between 0 and 1 and it is assumed by a researcher.
In our case I assume that the parameter should be chosen at a value of 0.5.

Then the result is:

Finally, I can compare the results of each method with a real volume of sales in 2010, which is equal to 2500:
I see that 3 years weighted average method reported 2225 - the closest value to 2500.