Bottled Iced
Tea Market
The purpose of this assignment was to engineer a
market share model and use that model to run Monte Carlo Simulation and predict
what is possible in the Iced Tea Market.
|
Inputs |
How Measured |
|
Sweetness
Market Share (initial) |
Percentage |
|
IceT
Market Share (initial) |
Percentage |
|
Small
Firm Universe Market Share (initial) |
Percentage |
|
Initial
Number of Small Firms |
Whole
Numbers > 0 |
|
Probability
that any Small Firm Will Exit Market |
Decimal |
|
Mean
Numbers New Entries Small Firms (Poisson
Distribution) |
Whole
Number |
|
Market
Share Parameters -
Sweetness -
IceT -
Small Firms |
Maximum
%, Minimum %, Most Likely % Representing
a Triangular Distribution |
|
Market
Share Parameters for Exiters -
To Sweetness -
To IceT |
Maximum
%, Minimum %, Most Likely % Representing
a Triangular Distribution |
The model was established to simulate the way that
market share flows among two dominant brands and the universe of entering and
exiting small firms in the Iced Tea Market.
In attempting to analyze this market, I created a
mathematical model to represent the way that the market share would change
based on the following flow diagram:
Based on the relationships from these variables, I
established the following:
Triangular distributions were used to model the
uncertainty in the market share changes per brand and small firm universe:
|
|
|
|
|
|
|
From Sweetness |
|
Minimum |
Most Likely |
Maximum |
|
To IceT |
|
1.00% |
5.00% |
10.00% |
|
To a Small Firm |
|
0.50% |
1.00% |
3.00% |
|
|
|
|
|
|
|
From IceT |
|
|
|
|
|
To Sweetness |
|
1.00% |
5.00% |
10.00% |
|
To a Small Firm |
|
0.50% |
1.00% |
3.00% |
|
|
|
|
|
|
|
From Small Companies |
|
|
|
|
|
To Sweetness |
|
5.00% |
10.00% |
15.00% |
|
To IceT |
|
5.00% |
10.00% |
15.00% |
Additionally, triangular distributions were used to
simulate the potential market share changes when firms exited the industry.
|
Exiting Small Firms Market Share (Remaining to IceT) |
|
|
||
|
|
|
Minimum |
Most Likely |
Maximum |
|
To Sweetness |
|
40.00% |
50.00% |
60.00% |
Finally, a binomial distribution was used to model
the uncertainty in the number of firms exiting the market in a given year.
The model was created to analyze the possible
changes in market share over the next ten years between the market’s two
dominant brands, Sweetness and IceT, as well as the universe of smaller
firms. As a result the models output was
generated as final market share for each brand category.
In order to determine the possible outcomes in
market share changes, I used the @Risk program to run
First, I created a DSS system whereby a simple
interface could be used to create the desired simulation output.
Next, I created inputs for @Risk by using probability distributions to describe the possible input
ranges. Finally, I used the @Risk
software to define the output ranges for the simulation model.
After defining the appropriate input and output, I
used @Risk to run one simulation constituting 1,000 iterations.
The following represent the output for year ten:
|
Outputs |
IceT /
Year Ten |
Sweetness
/ Year Ten |
|
Minimum |
0.433683544 |
0.389689684 |
|
Maximum |
0.602599919 |
0.554611742 |
|
Mean |
0.518785246 |
0.473865607 |
|
Standard
Deviation |
0.027446724 |
0.027112158 |
|
Variance |
0.000753323 |
0.000735069 |
|
Skewness |
-0.018484922 |
0.091005061 |
|
Kurtosis |
2.935168833 |
2.928519791 |
|
Number
of Errors |
0 |
0 |
|
Mode |
0.51455009 |
0.471901965 |
As the histograms and the descriptive statistics
indicate, the output cluster relatively normally
around the mean for each output range:
Sweetness (47.4%) and IceT (51.9%).
However there is a fair amount of variation in the output range for both
data sets (approximately 3%).
What does this information tell us? In simplest form it says that we can predict
the market share of the two dominant brands in ten years. These predictions are based on assumptions
about the market as a whole and about the entry and exit of small firms. The statistics indicate that within a 95%
confidence level, IceT will have a market share of between
49.1% and 54.5% and Sweetness will have a market share of 44.6% and 50.0%.
One improvement would be made in the data output
for the small firms that was rather irrelevant based on strange trends in the
output. I would like to take more time
to tinker with the model so that the entry and exit of small firms plays more
of a role in the modeling process.
Furthermore, over time it would be necessary for Sweetness and IceT to
validate and support the ranges used in the triangular distributions that
modeled the uncertainty in their market share changes. This would add more validity to the model.