Problem:

Demand Forecasting for Hi-End Smart Phones

Paul Jordan has just been hired as a management analyst at Hi-End Smart Phones, Inc. Hi-End manufactures a broad line of smart phones for the consumer market. Paul's boss, Douglas Levine, chief operations officer, has asked Paul to stop by his office this morning. After a brief exchange of pleasantries over a cup of coffee, he says he has a special assignment for Paul: "We've always just made an educated guess about how many we need to make each month. Usually we just look at how many we sold last month and plan to produce about the same number. This sometimes works fine. Bust most months we either have too many phones in inventory or we are out of stock. Neither situation is good."

Handing Paul the table shown here, Levine continues, "Here are the actual orders entered for the past 36 months. There are 144 phones per case. I was hoping that since you graduated recently from Montclair State University, you might have studied some techniques that would help us plan better. It's been a while since I was in college-I think I forgot most of the details I learned then. I'd like you to analyze these data and give me an idea of what our business will look like over the next 6 to 12 months. Do you think that you can handle this?"

"Of course", Paul replies. Two of our sales consultants have volunteered to help you out.

The first sales consultant, Jesse Rogers, suggested using the last month's demand as the best method for forecasting. He argued that smart phone customers are very fickle and that technology change so rapidly that this is the appropriate method. The other consultant, Hannah Balenty, argued that this method may not be satisfactory arguing that using the last period's demand as a predictor of the next period's demand produced erratic forecasts. For example, using this method the forecast would have predicted 480 cases for February 2009 and would have planned accordingly, but the demand was actually 436 cases. Clearly, this method could not sort out the fluctuations in the demand data, and it was therefore deemed unsatisfactory.

Hannah Balenty suggested using the average of all past demand to predict the next month's demand.   For May 2009, she would have predicted 461.5 cases [i.e., (480+436+482+448)/4], but 458 cases were actually demanded.   For the December 2009, she forecasted 468 cases (i.e., the sum of the eleven months demand divided by 11) and 525 cases occurred.    Jesse Rogers recognized that this averaging method produced forecasts that smoothed out the fluctuations but did not adequately respond to any growth or reduction in the demand trend. As a matter of fact, Mr. Rogers argued, the averaging method performed progressively worse as the amount of data increased. This was because each new piece of demand data had to be averaged with all the old data from month 1 to the present, and therefore each new element of data had less overall impact on the average. In fact, if BMI were to use the averaging method to forecast December 2011, the forecast would be 534.57142857142856 cases, clearly a poor forecast when compared with demand in the past few months.

Requirements - Prepare of Executive Summary that addresses these issues:

Required:

Question 1) Based on the data, propose a forecasting method for Hi-End to use to forecast the sales growth for smart phones. Be sure to explain why you choose the method that you did (HINT: you MUST plot your data). Forecast the demand for January 2012. Also, Use your model to predict the sales for smart phones for the remainder of 2012.

Question 2) Provide correct solution of the given problem with step by step calculations.