Jimboy A.

Ibañez BSAB-4

Case Problem 1: Forecasting Sales

The statistical Summary of the data is shown below:

Year January February March April May June July August September October November December Total Sales
1 242 235 232 178 184 140 145 152 110 130 152 206 2106

2 263 238 247 193 193 149 157 161 122 130 167 230 2250

3 282 255 265 205 210 160 166 174 126 148 173 235 2399

Total: 787 728 744 576 587 449 468 487 358 408 492 671 6755

Mean: 262.3333333 242.66667 248 192 195.6667 149.6667 156 162.3333 119.33333 136 164 223.6667 2251.666667

Variance: 400.3333333 116.33333 273 183 174.3333 100.3333 111 122.3333 69.333333 108 117 240.3333 21464.33333

StDev: 20.0083316 10.785793 16.52271 13.52775 13.20353 10.01665 10.53565 11.06044 8.326664 10.3923 10.81665 15.50269 146.5071102

The graph of the time series is as shown below.

The seasonal indices are calculated in the matrix below:
The following is a hypothesis test of seasonality:

Null Hypothesis: There is no significant seasonal component in the time series, versus

Alternative Hypothesis: There is a significant seasonal component in the time series

Conclusion: There is very strong evidence against the null hypothesis, hence the null hypothesis is rejected. There is
sufficient evidence to conclude that there exists significant seasonal component in the time series.

We now perform linear regression model with the following data:

Year Total Sales

1 2106
2 2250
3 2399

Linear Regression y = 146.5x + 1958.7

R2 = 0.9999
2450
Total Sales in 1000's --->

2400
2350
2300
2250
2200
2150
2100
2050
0 0.5 1 1.5 2 2.5 3 3.5
Year --->

This regression is extremely good with an R 2 value of 0.999. The regression equation is given as:

y = 146.5x + 1958.7

For the fourth year, the total sales are obtained by plugging in x = 4 in the above equation.
 y ( x  4)  146.5 * 4  1958.7  2544.7 (in thousands of dollars)

The average monthly sales during the fourth year, therefore, is 2544.7/12 = 212.058 (in thousands of dollars).

The forecast for a particular month (say July) is calculated by multiplying the average monthly sales forecast by
that month’s (July’s) seasonal index. For the month of July, it will be 0.831*212.058 = 176.22 (in thousands of
dollars).

The monthly forecasts for the 12 months of the fourth year are as shown below:

Month (yr.4) S. Index Forecast

January 1.398 296.45755

February 1.293 274.19143

March 1.322 280.3411
April 1.023 216.9357
May 1.043 221.1768
June 0.798 169.2226
July 0.831 176.2205
August 0.865 183.4305
September 0.636 134.8691
October 0.725 153.7423
November 0.874 185.339
December 1.192 252.7735

Suppose the actual January sales for the fourth year turn out to be $295,000. The forecasted January sales are $296,458.

Error between actual and forecasted sales = $296,458 - $295,000 = $1458

Error 1458
Percentage Error = *100  *100  0.49%
ActualSales 295000

This is an extremely small percentage error. Karen does not have to worry about this error and she can be assured that
her forecast model is extremely good.
Case Problem 2: AJ FORK AND HOE COMAPANY

Question 1: Comment on the forecasting system being used by AJ. Suggest changes or improvements that you believe
are justified.

 There are several weaknesses of current forecasting system:

1. Using only Qualitative analysis
 Forecasting figures are based on the meetings with managers
 No mathematical technique is involved
 Benefits: quick forecast and advantage of experience of each manager
 Demerits: forecast tend to be over inflated

Suggestion:
Implementation of quantitative method like seasonality technique with linear trend equation

2. Using actual shipment figure, instead of actual demand figures

 Marketing forecast is based on actual shipment data
 Trying to adjust for shortages in actual shipment data by anticipated promotions and environmental
and economical changes

Suggestions:
 Focus on past demand to project future demand
 Forecasting based on actual demand will help production department to schedule the
production line more effectively
 Provide a more clear picture to project realistic volume
 Create more sales and revenue for the company when anticipating the upward trend of
demand
 Prevent losses when anticipated downward trend in the market

3. Lack of communication between Production and Marketing department

 Both do not have accurate forecasting system and have different perception for the same
 Production department think that marketing department overinflates forecast
 Marketing generate unfaithful forecasts by adjusting part shipment and not predicting future
demands
 To maintain low-cost production , the long-term purchasing agreement is needed in order to keep
the price low for the raw material from suppliers, but having it just there is the price to pay for the
company

Suggestions

 Marketing should develop a forecasting system that reflects both past shortages and future
expected demands
 Meeting between the two should conduct at the end of each month
 Both departments should adjust the anticipated demand monthly to avoid unexpected
changes in the economy and shortage of the raw material
 4. Marketing division may not be optimistic

 Delay delivery problem was caused due to low productivity of production department.
 Current production is not sufficient to serve customer needs as it is based on “adjusted forecast”.
 Production capacity seemed not to be a problem as rake head & bow could be produced 7,000 &
5,000 units per day respectively, compared to the highest sales record in the last 4 years (month 11
year 1) at 83,269 units.
 Inappropriate inventory management was the major cause of unproductive production.

Question 2: Develop your own forecast for bow rakes for each month of the next year (year 5). Justifyyour forecast and the
method you used.

1. Naïve method
Naïve forecasts are the most cost-effective and efficient objective forecasting model, and provide a
benchmark against which more sophisticated models can be compared. For stable time series data,
this approach says that the forecast for any period equals the previous period's actual value.
2. Moving average method
 Moving average techniques forecast demand by calculating an average of actual demands from a
specified number of prior periods. Each new forecast drops the demand in the oldest period and replaces it
with the demand in the most recent period; thus, the data in the calculation "moves" over time

3. Weighted moving average method

When using a moving average method described before, each of the observations used to compute the
forecasted value is weighted equally. In certain cases, it might be beneficial to put more weight on the
observations that are closer to the time period being forecast. When this is done, this is known as a
weighted moving average technique. The weights in a weighted MA must sum to 1.
4. Additional seasonality model
5. Multiplicative seasonality model

Conclusion
•The table gives us the Four-Year Demand History for the Bow Rake and the demand figures are the number
of units promised for delivery each month. Hence we could not forecast using the exponential smoothing and
the trend adjusted exponential smoothing.
•There was no linear increasing or decreasing trend that was evident hence trend analysis for linear trends
had to be avoided.
•By analyzing the data provided we could observe a parabolic trend and seasonal variations with demand
increasing during the first 4 months and last 4 months.
•Using all the different techniques for forecasting and taking into considerations the error associated with
each we could conclude that the multiplicative model was the best forecasting technique.