Professional Documents
Culture Documents
Nofal Amin
Numrah Nadeem
Rida Naeem
Shaharyar Naeem
A-CAT FORECASTING CASE
• A-CAT Corp., a company that produces domestic electrical appliances in a poor region of
India.
• During the past several months, there has been an alarming dip in sales of its major product,
a voltage regulator that is used for varied purposes but most commonly as a protective
device for refrigerators and television sets,.
• The production department has been complaining about shortages of spares and components.
• Placing orders beyond a certain limit for the vital transformers used in most of its products
has also stretched the system - whereas the company previously had access to four suppliers
of transformers, now there is only one.
• The vice president has asked the chief operations manager to look into the problem.
VARIOUS METHODS AND
MODELS USED FOR
FORECASTING
NAÏVE METHOD
This method assumes that demand in the next period is equal to demand in the
most recent period.
Therefore, according to Naïve approach the forecasted demand in January 2011
would be 1053 units of voltage regulators which is equal to that of demand in
December 2010.
NAÏVE METHOD
As the name suggests it is a naïve approach, it won’t be 100% accurate but it’s a
good starting point since all the changes and information from the market have
been absorbed in December sales so the sales in January will not be very different
because the time horizon is not very long range.
Its also cost effective because we don’t have to go through complex methods to
calculate forecasted sales for next month.
EXPONENTIAL SMOOTHING
We have used two different values of alpha, i.e. 0.3 & 0.5. We have used two
different values just to make our work more accurate.
We have assumed our forecasted sales of January 2006 to be 779, i.e. same as of
January 2006 which we decided from the Naïve method.
When alpha is 0.3,
• MAD is 19.466
• MSE is 22734.392
• MAPE is 11.383
EXPONENTIAL SMOOTHING
The 3 month and 4 month moving averages calculated for A-Cat are shown in the
graph below:
WEIGHTED MOVING AVERAGE
For A-Cat the graph for 3 month weighted moving average is show below:
WEIGHTED MOVING AVERAGE
For A-Cat the graph for 4 month weighted moving average is show below:
WEIGHTED MOVING AVERAGE
• For 2011 January the forecasted voltage regulators for 3 month weighted
moving average turned out to be 1130 units and for 4 month weighted moving
average it turned out to be 1128 units which is quite close to the projected 4
month moving average units i.e. 1123 units
WEIGHTED MOVING AVERAGE
In this case two Regression/trend analysis were carried to see which would be
more beneficial in the future, the first one was between actual transformer sales
and period while period was the independent variable and transformer sales was
the dependent one.
The R square in the first regression was 0.405 which is really bad as the sales of
transformers is affected by only 40% by the period while it is affected by 60% by
other factors.
REGRESSION / TREND ANALYSIS
Seasonality is expressed in terms of the amount that actual values differ from
average values in the time series.
Analyzing data in monthly or quarterly terms usually makes it easy for a
statistician to spot seasonal patterns. Seasonal indices can then be developed by
several common methods.
It is a variable element in the time-series analysis of forecasting, and refers to the
phenomenon where the production and plan of product change on a certain
seasonal trend depending to the characteristics of the product.
SEASONAL VARIATION
The results of our seasonal variation will help us predict the sales of quarters in
future.
According to our result, the peak sales was in the second quarter, with a value of
1.245. The other 3 quarters had almost same sales with a slight difference.
BEST POSSIBLE METHOD
BEST METHOD SELECTED
The best possible method is regression. All other methods are not suitable because
their tracking signal is not in the range specified, i.e. -2.5 to 2.5. All values are
greater than this range.
We had the least error in 3 month weighted moving average, but the tracking
signals did not result in the range specified.
Then we applied tracking signal on 4 year weighted moving. Similarly, it was
applied on 3 and 4 month moving averages and exponential smoothing as well.
But the tracking signal result was not what we needed.
BEST METHOD SELECTED
While working on the case and analyzing it, we found out two qualitative factors
which can improve our forecasting.
Firstly, sales from composite method. In this approach, each salesperson estimates
what sales will be in his or her region. These forecasts are then reviewed to
ensure that they are realistic. Then they are combined at the district and national
levels to reach an overall forecast.
This will help us to overcome all the issues discussed in the case of the parts of
refrigerators and thus help increase our sales.
QUALITATIVE FACTORS
Secondly, we can use jury of execution opinion. Under this method, the opinions
of a group of high-level experts or managers, often in combination with statistical
models, are pooled to arrive at a group estimate of demand.
This will help us to forecast a better demand which will fulfill the needs of the
consumers and the company will not face any issues.
THANK YOU