Report on sales forecasts
Problem description
The company experiences a problem of great variability in sales as well as promotion and advertising expenditures from quarter to quarter. This makes it difficult to determine a sales forecast and prepare a budget for promotion and advertising expenditures.
Developing a multiple regression model
A multiple regression model can be used to determine the sales forecast given the values of the independent variables. In this case, sales of meatloaf mix is the dependent variable while the independent variables are promotion expenditure, advertising spending and economic index.
Description of the data
The data used in developing the model are historical data on each of the variables. They are time series secondary data collected by reviewing the company’s records. Quarterly sales data was obtained from the company’s sales reports or sales account while advertising and promotion expenditures were obtained from expenditure accounts or budget allocation reports.
Regression output results
Coefficient of determination
The model’s coefficient of determination as shown by the R-Square is 0.6059. This implies that that about 60.59% of the variations in quarterly sales were caused by changes in promotion expenditure, advertising expenditure and economic index (Curwin and Slater 380). The coefficient is more than 0.5 hence the model is good. However, the percentage variation explained by changes in variables outside the model is still high. The accuracy of the equation could be improved by incorporating such variables in the analysis.
The coefficient of correlation
The coefficient of correlation for the model as shown by Multiple R is 0.7783. This indicates that there is a strong correlation between quarterly sales and the three independent variables (Keller 136). Therefore, the three variables can be reliably used in determining quarterly sales.
Confidence level
The results indicate that the F Statistic for the model is 10.25 with a P-value of 0.000268073. Since the F statistic is more than two, the relationship between the variables is not accidental. Besides, the F Significance is less than 0.001 hence the model is statistically significant at 95% confidence level. Even at 99% confidence level, the model is still statistically significant.
Significance level for independent variables
The p-values for the coefficients of promotion and advertising expenditures are less than 0.05. This indicates that they are statistically significant at 95% confidence level. However, the coefficient of the economic index has a P-value of 0.1486. This is more than 0.05 hence the coefficient is not statistically significant at 95% confidence level.
Regression equation
Y = 883.524767 + 5.214703229X1 + 3.111679049X2 -5.628513903X3
Y is quarterly sales in thousands of dollars, X1 is promotion expenditure in thousands of dollars, X2 is advertising expenditure in thousands of dollars and X3 is the economic index. The equation indicates a positive relationship between sales and promotion and advertising expenditures a negative relationship between sales and economic index.
Forecast for the next two years
Next quarter
Advertising expenditure = $30,000
Promotion expenditure = $10,000
Economic index = 115
Sales, Y = 883.524767 + (5.214703229 × 10) + (3.111679049 × 30) – (5.628513903 × 115)
= 381.7430719
= $381, 743
Subsequent quarter
Advertising expenditure = $10,000
Promotion expenditure = $20,000
Economic index = 115
Sales, Y = 883.524767 + (5.214703229 × 25) + (3.111679049 × 10) – (5.628513903 × 115)
= 397.7300394
= $397,730
Suggestions going forward
The above tests indicate that the entire model is statistically significant. Besides, the coefficient of determination is more than 50%. However, the coefficient of the economic index is not statistically significant. To enhance the accuracy of the model, other variables affecting sales such as the price of meat loaf mix, prices of competing products, consumers’ income, among other variables should be added to the model. This will increase the coefficient of determination, among other measures, thus improving the model’s accuracy. The sales forecast obtained using the model as it is may be far from accurate.
Works cited
Curwin, Jon, and Roger Slater. Quantitative Methods For Business Decisions. London: Thomson Learning, 2007. Print.
Keller, Gerald. Statistics For Management And Economics, Abbreviated. Mason, OH: South-
Western Cengage Learning, 2015. Print
APPENDIX
