Wilkins, a Zurn Company: Demand Forecasting

The current demand forecasting method is based on qualitative techniques more than quantitative ones. If the forecast is not accurate, the company would carry both inventory and stock out costs. It might lose customers due to shortage of supply or carry additional holding costs due to excess production. If the actual demand doesn’t match the forecast ones, and the forecast was too high, this will result in high inventories, obsolescence, asset disposals, and increased carrying costs. When a forecast is too low, the customer resorts to a competitive product or retailer. A supplier could lose both sales and shelf space at that retail location forever if their predictions continue to be inaccurate. The tolerance level of the average consumer for product outages is quite low.
Forecasting for PVB:
Year PVB total Time D1 D2 D3
2001Q1 27512 1 1 0 0
2001Q2 45798 2 0 1 0
2001Q3 76968 3 0 0 1
2001Q4 43858 4 0 0 0
2002Q1 30580 5 1 0 0
2002Q2 53198 6 0 1 0
2002Q3 88704 7 0 0 1
2002Q4 51590 8 0 0 0
2003Q1 35372 9 1 0 0
2003Q2 57840 10 0 1 0
2003Q3 93388 11 0 0 1
2003Q4 58906 12 0 0 0
2004Q1 39382 13 1 0 0
2004Q2 75219 14 0 1 0
2004Q3 122868 15 0 0 1
2004Q4 54996 16 0 0 0

SUMMARY OUTPUT Regression Statistics
Multiple R 0.971266712
R Square 0.943359025
Adjusted R Square 0.922762307
Standard Error 7165.900459
Observations 16

ANOVA df SS MS F Significance F
Regression 4 9407636219 2.35E+09 45.80142 8.58623E-07
Residual 11 564851423.2 51350129
Total 15 9972487642 Coefficients Standard Error t Stat P-value Lower 95% Upper 95%
Intercept 32561.1875 5374.425344 6.058543 8.21E-05 20732.15708 44390.21792
Time 1977.63125 400.5860136 4.936845 0.000445 1095.947379 2859.315121
D1 -13193.1063 5207.618177 -2.53342 0.027803 -24654.99657 -1731.215929
D2 9631.5125 5130.004026 1.877486 0.087206 -1659.550226