Meriwell Case Forecasting
Meriwell Case Forecasting
Meriwell Case Forecasting
Develop a forecasting model using the regression method from your previous statistics class
with Prof Li.
SUMMARY
OUTPUT
Regression Statistics
Multiple R
0.891041167
R Square
0.793954361
Adjusted R Square
0.778969224
Standard Error
2534.130656
Observations
60
ANOVA
df
Regression
Residual
Total
4
55
59
Coefficients
SS
1.36E+09
3.53E+08
1.71E+09
Standard
Error
MS
F
3.4E+08 52.98279
6421818
t Stat
P-value
Intercept
Year
-1591333.333
800
Winter
-10466.66667
925.3337
-11.3112
5.63E-16
Spring
Summer
-12000
-8133.333333
925.3337
925.3337
-12.9683
-8.78962
2.3E-18
4.59E-12
Predicted Bags
3800
3800
2266.666667
2266.666667
2266.666667
6133.333333
6133.333333
6133.333333
14266.66667
14266.66667
Residuals
-1800
-800
733.3333
733.3333
1733.333
-133.333
866.6667
-133.333
-4266.67
-2266.67
RESIDUAL
OUTPUT
Observation
1
2
3
4
5
6
7
8
9
10
Significance
F
3.10865E-18
Lower 95%
Upper
95%
2522711.742 -659955
336.3974683 1263.603
12321.07679 -8612.26
13854.41013 -10145.6
-9987.74346 -6278.92
Lower
95.0%
-2522712
336.3975
-12321.1
-13854.4
-9987.74
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
14266.66667
3800
4600
4600
3066.666667
3066.666667
3066.666667
6933.333333
6933.333333
6933.333333
15066.66667
15066.66667
15066.66667
4600
5400
5400
3866.666667
3866.666667
3866.666667
7733.333333
7733.333333
7733.333333
15866.66667
15866.66667
15866.66667
5400
6200
6200
4666.666667
4666.666667
4666.666667
8533.333333
8533.333333
8533.333333
16666.66667
16666.66667
16666.66667
6200
7000
7000
5466.666667
5466.666667
5466.666667
9333.333333
9333.333333
9333.333333
-266.667
4200
-1600
-600
-66.6667
1933.333
1933.333
1066.667
-3933.33
1066.667
-3066.67
-3066.67
933.3333
5400
-3400
-400
1133.333
-866.667
133.3333
-1733.33
-733.333
2266.667
-866.667
-866.667
2133.333
2600
-1200
-2200
-666.667
-2666.67
333.3333
-1533.33
1466.667
5466.667
-666.667
-666.667
3333.333
5800
-2000
-5000
-2466.67
-3466.67
1533.333
-3333.33
-1333.33
666.6667
57
58
59
60
17466.66667
17466.66667
17466.66667
7000
2533.333
2533.333
4533.333
1000
In addition to forecasting demand of larger customers and aggregate demand, how might
the accuracy of the forecast be improved?.
The accuracy of the forecast could be improved by looking at demand of each month
specifically and also looking at particular peak weeks of the month. This will help in better
savings as there is seasonality with a month.
They could also try to segment the customers by geography as they serve a wide area and
the demand may be a function of particular geography.
4. What role should Ed Merriwell's "feel" of the market play in establishing new sales forecasts?
I think that the role of the feel of the market is unscientific way to establish new sales
forecasts. He should be able to enter the sales forecasting as accurately as possible in order
to do a good forecasts and improve his operating efficiency.
5. Would you recommend that Merriwell implement the regression method forecast?
Since Regression Analysis is one of the causal methods for demand projections and Meriwell
is struggling to forecast demand from its customers, Regression method is worthy of a try to
solve the problem.
But since Inventory control is also a issue, Time Series forecasting could also be used by
Meriwell Inc to predict their inventory and how better to do inventory management.
6. Suppose Merriwell was considering using Collaborative Planning, Forecasting, and
Replenishment (CPRF), what would you recommend to them?
The aim of CPRF model is to achieve more accurate forecasts. It is done by sharing information
in the supply chain with customers and suppliers. It compares forecasts down the supply chain
and looks for reasons in case on discrepancies. It works on consensus forecast. It works well
with B2B with few customers.
In Merriwell case, it has over 500 customers of various of varying order sizes. It will be difficult for
Merriwell to collaborate with its wide range of customers and do forecasts. It could first do a pilot
with its 100 largest customers by trade volume. It could then extend it to other large customers
but there is not going to any significant savings by investing in forecasts of small to very small
customers.