Electric Fan Industry - Case Study Example

Company analysis: Electric Fan Industrial Case Study Marie Jean Nova Southeastern Griffin Company analysis: Electric Fan Industry Case Study Executive SummaryThis report is about the analysis of an electric fan company, Bhagyanagar Fans Limited. The emergence of electric fans started in the 1900’s with India, the first to manufacture an electric fan, dominate the total production and the market share. Among the many fan companies in India, Hyderabad is considered the largest ceiling fan manufacturing center in the country.

Tibrewala, owner and chief executive officer of Bhagyanagar Fan Limited, was facing some shortages in the components supply in the last peak season, decided that a short term forecast for the next six months be done in order to deal better with this problem. Unfortunately, a relevant data could not be found and had to use a monthly production by organized sector since 2000 A.D. Anyway, this study will be analyzed to allow the manager to make a decision to choose which method is better in decision making.

The methods in question involve regression, and additive and multiplicative. After thorough scrutiny of all the methods, it was concluded that regression model was the most efficient method due to less errors. In order to reach this conclusion, a myriad of analysis tools such as, StatTools 6, MS Solver, were used in the analysis of sample data and results using spreadsheets. Background Information The Bhagyanagar Fans Limited Company has been experiencing a reduction in the components supply by many units in the last peak season.

So, the association decided to do something about it by hiring Ravi Kumar as a market research executive to come forward with a model that will stop the company from losing money and help them better understand the short term demand pattern. But according to David W. Stockburger, in order to make this forecast, or the info which needs to be predicted, must be obtained from some kind of sample data, then transform this information into the predicted. The young executive, Ravi Kumar, despite his unsuccessful research data, was accepting the challenge to work with the only available data he could find, the monthly production from 2000 A.D., in order to come up with a short term forecast for the next six months as requested by the owner, MR.

Tibrewala. Different tools were used in the analysis of the electric fans production such as, MS Excel Solver, and StatTools 6 to determine the best forecasting model where the properties of regression, additive and multiplicative models were discussed. After all of these models were analyzed using the historical data, the trend and the seasonality which had the most impact on the data. The regression model was chosen as the most appropriate forecast based on the RMSE, using the rule of error, “the smaller the better” (Makridakis et all, 1998).

Problem Will Ravi Kumar, the young research executive, be able to use the monthly production from 2000 A.D. data instead of a relevant data on sales or demand for electric fans and still generate an appropriate forecast for the next six months? AnalysisThe first method used was the regression model where the historical data were computed as dependent variables, the time and the seasonality as independent variables.

For the time, the increasing numbers were used from 1 to 111 including the forecast for half of the year ahead. For the seasonally, the dummies were used to map each month and using 1or 0 (Albright, Winston & Zappe, 2010). Then to figure out the correlations between these independent and dependent variables of the historical data, the regression model was used. After the report of the regression statistics summary had been carefully analyzed, it was evident that the behavior of that historical data, the R-square (0.70641) was more than 70% of these dependent variables.

In addition, the multiple R was 0.84048 which showed that the fitted values were very close to the original data.Based on the significance F of the regression, the model was proven to be good even with the 99% confidential level. Furthermore, based on the p- value and the t statistics, all independent variables were significant (t>2 and p< 0.01). Therefore, the forecast model was built using the multiple linear regression equation, F= a+ b t+ B1Q1+b2Q2+…. (Albright, Winston & Zappe, 2010) and the following coefficients from the regression summary output were used: CoefficientsIntercept6.

063752735Time0.032333453Jan1.029907707Feb1.322574254Mar1.952740801Apr1.151908066May1.519574613Jun1.020574493After the forecast model was analyzed, it was checked out by calculating the error statistics. The “e” and the “e^2” were computed as well as the absolute percentage errors, then the Mean Error (ME) and the Root Mean Squared Error (RMSE) were calculated from the calculated error metrics (Makridakis, Wheelwright & Hyndman, 1998). After that, the Thiel’s U statistic was also computed to know the efficacy of the model.

“A value of U>1 indicates a poor forecasting model relative to a naïve forecast. A good forecasting model has a value of up