You are on page 1of 6

Monte Carlo Simulation in Excel

1. Fundamental Excel Functions

RAND() Function

RAND() function is not particularly designed for the Monte Carlo simulation; however, this Excel function
can generate a random number that is between 0 and 1 (inclusive). Thus, we can use RAND() function to
generate a random number Δ for an unknown factor in one scenario. We then use Δ as a “pseudo
probability” of the unknown factor, in order to find the most likely value for the unknown factor in the
scenario.

NORM. INV() Function

As I demonstrated in the lecture class, the NORM.INV() function is to find the corresponding value
according to a given probability. When we use RAND() to generate a Δ value, we can then give this Δ
value to NORM.INV() to find the most likely value for the unknown factor that follows a normal
distribution.

2. Monte Carlo Simulation – Finding the Probability Distribution for Each Unknown Factor

Before the Monte Carlo simulation, we must identify the unknown factors and the probability
distribution for each unknown factor. Usually, the probability distributions are given to us. We only need
to build these distributions in the data set.

For the uniform distribution, we need to record the lower bound and upper bound. The part cost per
unit in our case follows a uniform distribution. The part cost per unit can be any value between $80 and
$100. Thus, in cell F14, we type 80. In cell F15, we type 100.

For the normal distribution, we need to record the mean and the standard deviation. The first year
demand in our case follows a normal distribution. Thus, in cell F18, we type 15000. In cell F19, we type
4500, as shown in the picture below.
For the discrete distribution, the given probabilities are the ranges of probability for each possible
outcome. We then need to build a probability interval for each discrete outcome to reflect this range. In
our case, we are given the following distribution for the direct labor cost per unit.

Direct Labor Cost Probability Probability Interval

[0, 01)
$43 0.1
The length of probabilities is 0.1, as given on the right.

[0.1, 0.3)
$44 0.2
The length of probabilities is 0.2, as given on the right.

[0.3, 0.7)
$45 0.4
The length of probabilities is 0.3, as given on the right.

[0.7, 0.9)
$46 0.2
The length of probabilities is 0.2, as given on the right.

[0.9, 1)
$47 0.1
The length of probabilities is 0.1, as given on the right.

For instance, if the software generates 0.4 as the value of Δ for the direct labor cost, we look up 0.4 in
the probability intervals. It lies in the [0.3, 0.7) range, which points to $45. Then, we use $45 as the
estimated value for the direct labor cost in the scenario.

If the software generates 0.82 as the value of Δ for the direct labor cost, we look up 0.82 in the
probability intervals. It lies in the [0.7, 0.9) range, which points to $46. Then, we use $46 as the
estimated value for the direct labor cost in the scenario.
In Excel, from cell A13 to cell D19, we implement the probability intervals as shown in the picture below.
(I have built the intervals in the source data set. So you don’t need to build it again.)

3. Monte Carlo Simulation – Build the Profit Equation

In the range A4 to B11, we build the profit calculation equation and factors being used in the equation,
as shown in the picture below. As you can see in cell B6, B7, and B8, we use RAND() function to generate
the random numbers for these unknown factors. Excel will then look up the probability distributions we
established in step 2 to find the estimated value for each unknown factor in a scenario according to the
probability distribution, and then use the estimated values to calculate the predicted profit in that
scenario.
4. Monte Carlo Simulation – Build Trials

In the range A22 to E1021, we build the 1000 hypothetical scenarios. To simplify the range selection, I
hid the scenarios from scenario 22 to scenario 999.

In cell B22, we type =B6 to refer to the estimated direct labor cost per unit. In cell C22, we type =B7 to
refer to the estimated part cost per unit. In cell D22, we type =B8 to refer to the estimated first year
demand. In cell E22, we type =B11 to refer to the profit.

Select the range A22 to E1021, and then click the Data tab on the top, and then select the Data Table
analysis under What-if Analysis
In the Data Table window, click on cell D1 for the Column input cell option, and then click OK.

We can find that the range A22 to E1021 is filled by different possible values in different scenarios.

At the same time, we received the probability distribution of the predicted profits, as shown below:

Profit Distribution
200

150

100

50

0
We also received some other insights about this new product development, such as the probability of
losing money for selling this new smartphone (0.077), what is the possibly minimum predicted profit in
our analysis (-$871104, losing money in this case), what is the possibly maximum predicted profit in our
analysis ($2689520, receiving profits in this case), as shown below.

Why Monte Carlo Simulation?

The market can be highly complicated. The customer’s demand for a new product is difficult to predict
as well. Currently, if the analyst only predicts that there would be only 3 scenarios as I introduced in the
video lecture, the prediction perspective is very narrow. The results from the analysis may not be
reliable and accurate.

However, if we increase the possible scenarios to 1000 by using the Monte Carlo simulation, we can
“simulate” the unknown future with various possibilities in the analysis. We will be better informed
about the possible outcomes when making a decision.

You might also like