SMU EMIS 4395 - Developing a Forecasting Model

Unformatted text preview:

Page 1Page 2Page 3Page 4Page 5Page 6Page 7Page 8Page 9Page 10Page 11Page 12Page 13Page 14Page 15Page 16Page 17Page 18Page 192002-02 Spring 2002 Forecasting for the Future:Developing a Forecasting Modelfor Brinker International Latonya Morris, James On •Forecasting for the Future: Developing a Forecasting Model for Brinker InternationalDeveloped by:Latonya Morris and James OrrEMIS 4395May 7, 2002ioManagement Summary Brinker International is one of the leading trendsetters in the restaurant industry. The company started in 1975 with just the Chili's restaurants. Under the direction of Norman Brinker the company was renamed to Brinker International Inc. The Brinker International family now includes Chili's, Corner Bakery, Cozymel's, Maggiano's, Big Bowl, Eat Zi's, Rockfish, Macaroni Grill, and On The Border. The problem consisted of forecasting sales for Brinker International, more specifically The Corner Bakery. To obtain a more accurate forecast a time series decomposition model was used. The multiplicative model consisted of four parts: trend, seasonal, cyclical, and irregular component. All of these components were solved for using historical data, except for the irregular component, which was not computed. The model is designed using excel in a manner that allows employees to enter the data as it develops and update the forecast automatically. Using the trend, seasonal, and cyclical lines we were able to obtain a very accurate estimate of what the sales are going to be for the year 2003. The forecast was slightly limited due to a lack of data. As more data becomes available the model will become more accurate. The sales forecast for 2003 currently stands at $9,926,506.10. This forecast does not take into consideration the increase in revenue due to new stores opening.Background and Description of the Problem Situation Brinker International, founded by Norman Brinker, is a company that started in 1975 with the one original Chili's location. Chili's is still apart of Brinker International family, but seven other restaurants join it. The Brinker International family now includes Chili's, Corner Bakery, Cozymel's, Maggiano's, Big Bowl, Eat Zi's, Rockfish, Macaroni Grill, and On The Border. Due to time constraints we were only able to focus on the problems of the Corner Bakery. The problem that Corner Bakery faces is that there is not an accurate model to forecast sales. In order to overcome this problem time series decomposition model was used. The multiplicative model consisted of four parts: trend, seasonal, cyclical, and irregular component. All of these components were solved for using historical data, except for the irregular component, which was not computed lack of data. Our goal is to use excel to produce an updateable model to forecast sales for the following years. The forecasting model is tested against the actual sales for each year to test the accuracy of the model. The forecast will allow Corner Bakery to have a very good estimate of the sales that they can expect for each of the months for 2003.LIAnalysis of the Situation •The problem consisted of two parts: the developing an accurate forecasting of sales data for Brinker's 2003 fiscal year, and second, finding a way to put our forecasting model in a Microsoft Excel spreadsheet. The forecast was requested to be incorporated into a Microsoft Excel spreadsheet for the ease of our client's use. To forecast sales many different forecasting models were tried: first, exponential smoothing and moving average models were used to forecast sales. These methods were found to be unreliable and not very accurate. Finally, a multiplicative time series decomposition model was used to forecast sales. This model was found to be the most accurate and the most beneficial to •our client's situation. The model that was used for the forecast was Y1 = x Si x C1 x I. The multiplicative time series decomposition model consists of four parts: a trend component, seasonal component, cyclical component, and an irregular component. The forecast for month one would be Y1 = T1 x S1 x C1 x Ii. •Technical Description of the Model The model that was used to forecast the following year's sale is a multiplicative time series decomposition model. The model takes the form Y1 = Ti x C1 x Si x I. Ti is the trend component. The C1 is the cyclical adjustment index. Sirepresents seasonal adjustment index and Ii is the irregular, random index. To compute for the Ti first the seasonal index must be found. To find the seasonal index a twelve-month moving average was taken over the available sales data. The first average was taken from the first to the twelfth month and the second average was taken •from the second to the thirteenth month. This process is repeated until all the months of sales data have been incorporated. From these twelve-month moving averages the centered moving averages can be computed. The first centered moving average is the average of the first and second twelve-month averages. This first centered moving average corresponds to the seventh period of sales data. •To obtain the Si for that month divide the sales data by centered moving average for that corresponding month. To get a more accurate seasonal index, the seasonal indexes are grouped by month. Once all the seasonal indexes from the same month are grouped together, an average of the indexes for that month are taken. The average for each individual month is summed with the other months and if the sum does not equal twelve, the indexes must be adjusted. The adjustment is made by taking the average seasonal index for each month, multiplying it by twelve and then dividing that number by the sum of the original indexes (i.e. [12 * Sj]/ Sum of original indexes). •Each individual Ti is just the trend data values that move through the deseasonalized sales data. The trend line was found by using the number of months of sales data and the deseasonalized sales data for that month. The deseasonalized sales data is found, by multiplying the actual sales data by the adjusted seasonal index. The cyclical component for each month was found by taking the centered moving average for that month divided by the trend forecast for that month. Then the cyclical components were adjusted in the same manner the seasonal indexes were adjusted. The irregular random index could not be computed and therefore was not integrated into our forecast. The data for


View Full Document

SMU EMIS 4395 - Developing a Forecasting Model

Download Developing a Forecasting Model
Our administrator received your request to download this document. We will send you the file to your email shortly.
Loading Unlocking...
Login

Join to view Developing a Forecasting Model and access 3M+ class-specific study document.

or
We will never post anything without your permission.
Don't have an account?
Sign Up

Join to view Developing a Forecasting Model 2 2 and access 3M+ class-specific study document.

or

By creating an account you agree to our Privacy Policy and Terms Of Use

Already a member?