Prepared by: Karriem Pierre
November 20, 2011
Nova Southeastern University
Business Modeling, Fall Semester, Online Course
Professor Phillip S. Rokicki, Ph.D
Ms. Quintana CEO of Northern Napa Valley Winery Inc. was considering conducting business with Trans Continental stores to sell excess grapes from the 2008 harvest. Prior to making a decision Quintana must determine how much of the harvest should be retained for the production of Northern Napa’s own red table wine. Quintana realized that the quantity of red table wine produced is closely associated to the sales. Ms. Quintana reviewed her past sales in order to predict future sales based on the amount of wine produced in 2008. The issue Quintana faced was the seasonality of wine sales; December being its peak time in sales. After careful consideration, Quintana determined that although her previous forecasting method had worked from an operational standpoint, however, there was still room to improve its accuracy. Quintana has to determine what the forecasting profits will be for the next 16 months (Sept2008 to Dec 2009) prior to the conclusion of any business deals with TransContinental. The software that was used in order to predict the future sales was StatTools. StatTools provided me with four different methods (Moving Averages, Simple, Holt’s & Winters’ exponential smoothing) of forecasting to identify the most accurate prediction possible. With a Means Absolute Percentage Error (MAPE) of 3.72% the Holt’s exponential smoothing method was determined to be the best method of predicting monthly wine sales for the next sixteen months (Figure 2).
Background/Additional Research on Napa Valley Wineries
The Napa valley is known for its stylish and exquisite wine blends. The highly acclaimed Paraduxx Winery was founded in 1994 and its current wine maker is David Marchesi. Paraduxx is the only winery that is truly devoted to wine blends in the Napa Valley. Paraduxx has four different estates that produce grapes with distinct attributes that reflect the variations in soil, terrain, microclimate, and exposure. This vineyard has created a blend of zinfandel, Cabernet Sauvignon, along with smaller amounts of Merlot and Petite Sirah to create the first Paraduxx vintage. One of Parduxx’s quality wines is the 2006 Postmark Monitor Ledge Vineyard Red Wine. This blend of Zinfandel paired with Cabernet Sauvignon and Merlot originates solely from our Monitor Ledge Vineyard which is known for producing concentrated and structured wines. This wine highlights red cherry and rose petal aromas, with notes of blueberry, blackberry, smoky cedar and cola. The flavors of the red fruit and warm blackberry with a dash of cinnamon spice. Analyzing the Problem
The data used was gathered from previous sales from Jan. 2000 through Aug. 2008. This data will be used in forecasting sixteen month time period starting with Sept. 2008 and all of 2009 wine sales. The forecasting tool used was StatTools, which helped in determining the total wine sales for the next sixteen months. The question that will be addressed: What is Northern Napa Valley Winery’s projected sales for the upcoming sixteen months? To better answer this question there are four different methods of forecasting (Figure 5 Moving Averages, Figure 4 Simple, Figure2 Holt’s & Figure 3 Winters’ Exponential Smoothing) that will be used, of those four one will give the best prediction based on the past data. Forecasting methods where done deseasonalized. The Mean Absolute Percentage Error is important when forecasting data. The closer the MAPE is to zero the more accurate the forecast. This will be the catalyst in determining the accuracy of the forecasts. Runs test (Null Hypothesis of Randomness)
Prior to doing any forecasting it...