Professional Documents
Culture Documents
: Forcasting
Group No : 8
Particpant Name
# Roll No Name
1 2023EXEMBA03 Anupam Mandal
2 2023EXEMBA08 Bablu Kumar Patel
3 2023EXEMBA14 Hemanta Kumar Mardi
4 2023EXEMBA30 Shakti Nanda
5 2023EXEMBA32 Siddharth Shankar Pujari
Case Summary
Company Background
Case about How A Cat Copr started business in 1986 who is a leading producer of electric producer
in India, medium scale industry, mainly focused on Rural population of Vidarba Region.
It has alliance in business firm like Juptiar Inc & Global Venture. Employee Strength is 40 Nos. Primary
functional dept -Purchase, Design, Manufacturing and Sale Dept.
Its primary flagship product a voltage regulator of 500 KVA (Brand VR 500)
Issues
in summary - improper forecasting of demand and poor inventory mangement resulting sluggish growth
Assignment Objectives:
Analysis of reference data, forecasting the next order of vital item like trasnformer by applying different forecasting
method. Identifying the errors and recommendations based on the analysis
ent forecasting
Exibit-1
Transformer Requirements during this period ( Taken from the sale of Voltage Regulator)
Exibit-2
Sales figures of Refrigetor
quarter 2006 2007 2008 2009 2010
1 3832 4007 4826 5411 6290
2 5032 5903 6492 7678 8332
3 3947 4274 4765 5774 6107
4 3291 3692 4972 6007 6729
Historical Data Analysis
1600
Period 2006 2007 2008 2009 2010
Jan 779 845 857 917 887 1400
916
902
898
888
818
802
784
779
716
708
708
695
Observations
Seasonal Variations
J an Feb Mar Ap r May Jun Jul Au g Sep Oc t Nov D ec
Peak Seasons: The months of April through July are the peak demand seasons, which could
be linked
to the onset of summer and higher use of voltage-dependent appliances Trasnformer ReQ-2007
1133
1124
1056
927
889
871
857
845
820
772
751
739
Low Seasons: Demand significantly drops in August, possibly due to monsoon season in
India, which might reduce the usage or purchase of electrical appliances.
Lowest demand is seen at the beginning and end of the year, with January , Feb & Sep
J an Feb Mar Ap r May Jun Jul Au g Sep Oc t Nov D ec
Trasnformer ReQ-2008
1246
1198
1159
Year-on-Year Growth
1072
990
945
937
922
881
879
857
798
The year-on-year growth rate in demand shows a declining trend in growth percentages,
although the demand itself is increasing.
The growth from 2006 to 2007 was approximately 12%. J an Feb Mar Ap r May Jun Jul Au g Sep Oc t Nov D ec
The growth from 2009 to 2010 slowed down to approximately 2%. However, in last five year
growth incresed from 9614 to 13214
Trasnformer ReQ-2009
1356
1288
1276
1142
1100
1082
1009
1001
998
956
917
877
Year-on-Year Growth
1400 0.14 J an Feb Mar Ap r May Jun Jul Au g Sep Oc t Nov D ec
1200 0.12
1000 0.1
800 0.08 Trasnformer ReQ-2010
1297
1256
1202
1197
1170
1163
1118
600 0.06
1053
997
982
400 0.04
892
887
200 0.02
0 0
Growth %
J an Feb Mar Ap r May Jun Jul Au g Sep Oc t Nov D ec
Recommendations
Increase Inventory Before Peak Season: A-CAT Corp. should consider increasing its transformer inventory before the peak season starts (around February to March) to
accommodate the increased demand without facing stockouts.
Review and Adjust Post-Peak Inventory: After the peak season, especially post-July, inventory levels should be carefully adjusted to prevent overstocking, considering the
significant drop in demand in August
The current forecasting methods used by A-CAT and their limitations
The company was estimating the number of transformers needed based on the sales figures of the last two to three
combined with the sales figures from the same months in the previous two years. This method appears to blend ele
series analysis (looking at recent and seasonal past data) with a form of qualitative judgment (estimating future nee
recent past sales).
It potentially overlooks deeper insights that could be gleaned from the data, such as cyclic trends, outlier impacts, o
growth deceleration observed in recent years
The method seems to lead to situations where the company either ends up with excess inventory (overstocking) or
meet production needs (stockouts)
es of the last two to three months
thod appears to blend elements of time
ent (estimating future needs based on
A) Naive Approach
In a naive approach, the forecast for the next period is simply the actual demand of the current period. This method assumes
Obervations
Variability in Error: The percentage errors vary significantly across different months and years, indicating fluctuations in the ac
capture.
Certain months, like March and May, show relatively small errors across the years, suggesting some stability in demand patter
unpredictability.
Trend Over Years: While the average error in 2008 is 7%, it slightly decreases in 2009 to 6%, and then significantly increases in
unpredictability in demand or changes in market conditions not reflected in past data
riod. This method assumes that what happened in the current period is the best predictor of what will happen in the next period
Forecast
for 2011 Naive Approach Error
cating fluctuations in the accuracy of the naive forecasting approach. This suggests that demand for the product is subject to factors that
e stability in demand patterns during these periods. Conversely, months like October and November exhibit larger errors, especially in 201
en significantly increases in 2010 to 12%. This upward trend in error percentage suggests that the naive method's accuracy may be deteri
pen in the next period
oduct is subject to factors that this simple forecasting method does not
Forecast of SMA
Obervations
There is significant variability in the percentage errors, with some months showing errors as low as -1% and others as high as -
term trends, it may still struggle with abrupt changes in demand or exceptional events
Forecasting Error % Error
s -1% and others as high as -46%. This variability suggests that while SMA can smooth short-term fluctuations and highlight longer-
% Error
2010
-2%
-4%
-6%
-4%
-3%
-1%
-2%
-21%
-16%
-46%
-25%
-13%
5%
d highlight longer-
Applying different forecasting techniques
Forecast
Obervations
Seems fair in peak season where as Oct- Dec its showing high variability
Forecasting Error
2009 2010
-9% 0%
-17% 0%
-12% -4%
-10% -1%
-20% -1%
-18% 2%
-17% 1%
-24% -17%
-10% -16%
-24% -40%
-31% -17%
-11% -9%
1% 7%
Applying different forecasting techniques
D) Exponential Smoothing
It is a weighted average method that includes all past data in the forecasting calculation. The smoothing constant (α) can
be changed to give more weight
to recent dataAt-4
when the value
At-3 is At-2 At-1 At Fore
high or more weight to past data
Period
when it is low 2006 2007 2008 2009 2010 2007
Ft-1 of
At-1 Ft-4
2006
Jan 779 845 857 917 887 779 779 779
Feb 802 739 881 956 892 802 802 802
Mar 818 871 937 1001 997 818 818 818
Apr 888 927 1159 1142 1118 888 888 888
May 898 1133 1072 1276 1197 898 898 898
Jun 902 1124 1246 1356 1256 902 902 902
Jul 916 1056 1198 1288 1202 916 916 916
Aug 708 889 922 1082 1170 708 708 708
Sep 695 857 798 877 982 695 695 695
Oct 708 772 879 1009 1297 708 708 708
Nov 716 751 945 1100 1163 716 716 716
Dec 784 820 990 998 1053 784 784 784
11884 13002 13214 yearly 11621
Obervations
Seems fair in peak season where as Oct- Dec its showing high variability
Say Alpha 0.5
Considering Forecasted Sale of 2007-1=2006 is equal to ac
D) Exponential Smoothing
It is a weighted average method that includes all past data in the forecasting calculation. The smoothing constant (α) can
be changed to give more weight
to recent dataAt-4
when the value
At-3 is At-2 At-1 At Fore
high or more weight to past data
Period
when it is low 2006 2007 2008 2009 2010 2007
Ft-1 of
At-1 Ft-4
2006
Obervations
Seems fair in peak season where as Oct- Dec its showing high variability
Say Alpha 0.5
Considering Forecasted Sale of 2007-1=2006 is equal to a
Forecast
2007 2008 2009
Obervations
Seems fair in peak season where as Oct- Dec its showing high variability
0.5
6/1/2007 1124 0
5/28/05 10/10/06 2/22/08
7/1/2007 1056
8/1/2007 889
9/1/2007 857
10/1/2007 772
11/1/2007 751
12/1/2007 820
1/1/2008 857
2/1/2008 881
3/1/2008 937
4/1/2008 1159
5/1/2008 1072
6/1/2008 1246
7/1/2008 1198
8/1/2008 922
9/1/2008 798
10/1/2008 879
11/1/2008 945
12/1/2008 990
1/1/2009 917
2/1/2009 956
3/1/2009 1001
4/1/2009 1142
5/1/2009 1276
6/1/2009 1356
7/1/2009 1288
8/1/2009 1082
9/1/2009 877
10/1/2009 1009
11/1/2009 1100
12/1/2009 998
1/1/2010 887
2/1/2010 892
3/1/2010 997
4/1/2010 1118
5/1/2010 1197
6/1/2010 1256
7/1/2010 1202
8/1/2010 1170
9/1/2010 982
10/1/2010 1297
11/1/2010 1163
12/1/2010 1053
1/1/2011 -71950.73
2/1/2011
3/1/2011
4/1/2011
5/1/2011
6/1/2011
7/1/2011
8/1/2011
9/1/2011
10/1/2011
11/1/2011
12/1/2011
Actual Sale
82363