FINANCIAL

MODELLING
STEPS IN FINANCIAL MODELLING
STEP 1 Define & Structure the problem

STEP 2 Define the Input and Output Variables of the Model

STEP 3 Decide Who Will Use the Model and How Often

STEP 4 Understand the Financial and Mathematical Aspects of the Model

STEP 5 Design the Model

STEP 6 Create the Spreadsheets

STEP 7 Test the Model

STEP 8 Protect the Model

STEP 9 Document the Model

 HOW WILL WE PROCEED WITH MODELLING?

1. Insert historical data.

2. Analyze historical data by computing ratios, growth
rate etc.

3. INPUT: Making assumptions for the future.

For assumptions, we use:

• Management Inputs
• Historical Trend
• Market Projections
4. Working Schedules
• Revenue workings
• Operating cost workings
• Fixed Assets schedule
• Debt schedule
• Any other working schedule

5. OUTPUT of the model

• Projected P&L
• Projected B&S
• Projected CFS

6. Use the above output for Business Valuation

 MAKING
PROJECTIONS
 PROJECTING REVENUE
Revenue can be quite difficult to predict. How will we know what
revenue will be in the next year? The truth is, it is almost impossible to
be 100 percent sure. We will need to make an assumption with the
understanding that assumption will come with a degree of uncertainty,
and may therefore change.

So how can you best make rational predictions for the future?
It is important to research and understand the company’s business
model, gathering as much information as you can to make your own
best judgment. Revenue is almost always driven by a product of
pricing and volume.
 PROJECTING REVENUE
When thinking about projecting revenue, your research should focus
on understanding the company’s pricing and volume.

• What initiatives is the company taking to increase its volume in the

upcoming year?
• Is it increasing its advertising?
• Is it acquiring other businesses or customers?
• What outside forces could affect the company’s pricing model?
• Is it increasing its prices?
• Is it facing tremendous market competition and must lower its
prices?
 PROJECTING REVENUE
Alternatively, revenue can be forecasted using the following formula:

NY Total Revenue = CY Total Revenue × (1 + NY Revenue Growth Assumption)

NY=Next Year, CY=Current Year

Historical trends can help us determine how best to make initial

projections, with the knowledge that we can later tweak as we build a
more fundamental understanding of the business.
 SOURCES OF INFORMATION FOR
MAKING PROJECTIONS
1. Investor presentations: Try to look for a recent investor
presentation on the investor relations section of the company web
site. These presentations are typically designed to explain recent and
future performance to existing or future investors of the company’s
stock. These presentations can contain high-level projections.

2. Earnings calls: One can easily find when the next earnings call is on
the investor relations section of the web site. At the earnings call, you
can listen to the management speak about the company’s most recent
financial performance. Management also sometimes gives guidance
on the company’s future performance.
 SOURCES OF INFORMATION FOR
MAKING PROJECTIONS
3. Research reports: If you can get your hands on an equity research
report or a credit rating report by analysts who has followed the
company for several years, that report would contain estimated future
performance.

4. Data sources: Yahoo! Finance, Bloomberg etc. are examples of data

sources that contain market consensus estimates. Yahoo! Finance is a
free resource, so, if you do not have access to a paid service, this can
serve as a good reference.
 PROJECTING COSTS
First, it is important to consider whether the costs are fixed or variable.

A fixed cost is relatively static and may grow a certain percentage year
over year.

For example, rent can be considered a fixed cost as it may only

increase 5–10 percent each year, independent of the growth in
revenue.
 PROJECTING COSTS
In contrast, a variable cost will increase in direct proportion to the
growth of the business, most commonly determined by the revenue
growth.

If the revenue is increasing by 10 percent, the costs will also increase

by 10 percent.

If the revenue decreases by 4 percent, the costs will also decrease by

4 percent.
 PROJECTING COSTS
Quite often cost of goods sold is considered a variable cost. If your
revenue is declining, you are most likely selling less product, so your
costs should also be decreasing.

Conversely, if your revenue is increasing, you are most likely selling

more product, so cost of goods sold should be increasing in direct
proportion to the revenue.
 FORECASTING COGS
COGS & Operating Expenses can be projected using the following
equation:

NY COGS = NY COGS as a % of Revenue x NY Total Revenue

 Depreciation and Interest
When building a complete financial model it is recommended to leave
projected depreciation and interest expense empty for once.

We will build a depreciation schedule that will contain projected

depreciation expense to be linked in there.

Interest figures will be taken from the Debt schedule.

 FORECASTING METHODS
1. Conservative (the minimum of the past three years)
2. Aggressive (the maximum of the past three years)
3. Average (the average of the past three years)
4. Last year (recent performance)
5. Repeat the cycle
6. Year-over-year growth
7. Project out as a percentage of an income statement or balance
sheet line item
 CONSERVATIVE
• Generally, we assume money spent is more conservative than
money received.

• So, taking the minimum amount from the last three years may not
be the most accurate, but it is a conservative approach.

• You can use the “minimum” formula in Excel. For example,

“=min(x,y,z)” will give you the lowest amount of x, y, and z.
 AGGRESSIVE
• This is probably not the most recommended method, but it is a
possible method, so we will note it.

• Assuming more money received is more aggressive, we would take

the maximum amount from the last three years.

• You can use the “maximum” formula in Excel; “=max(x,y,z)” will

give you the maximum amount of x, y, and z.
 AVERAGE
This is a popular method, but be warned that quite often the average
of the past three years does not always give the best indication of next
year’s performance, especially if one of the past three years was
unusual.

Many analysts consider using the average method as the safety

method. We recommend better to carefully go through all various
methods before considering the average method.

You can use the “average” formula in Excel; “=average(x,y,z)” will give
you the average amount of x, y, and z.
 LATEST YEAR
This is based on the underlying assumption that the company’s
performance last year is most indicative of its future performance. If
one does not know the business or the specific line item well, it may
not be easy to determine if this is correct method to use.

However, a combination of this method and the conservative method

is a quite useful indicator. In other words, if last year’s performance
also happens to be the most conservative of the last three years, then
we have two supporting methodologies that point to the same number.
The more support we have, the better.
 REPEAT THE CYCLE
Quite often the last three years’ numbers will be quite volatile,
swinging from positive to negative or from a very small value to a
large value.

Although it is often difficult to identify exactly why, some companies

can plan more significant events every second or third year.

For example, companies can make larger capital expenditure

investments every third year, and smaller investments in the other
years. In this case, you may want to continue this trend. The easiest
way to do this is to have the projected year to equal to the first
historical year.
 YEAR-OVER-YEAR GROWTH
Here we can assume some year-over-year growth rate to project the
line item going forward. The growth rate can be dependent on what
exactly that “other” line item is.

If it is rent, for example, we can assume the rent will increase by a

standard 5 percent each year.

You can also take a look at the historical trends and apply those
trends to the projections.
 As a % of an income statement or
balance sheet line item
Some of the line items can sometimes grow dependent on another
income statement or balance sheet line item.

For example, in case of employee salaries, you may want to project

this line item based on a percentage of SG&A. One way to determine if
this can be an appropriate method is by looking at the historical
percentage of SG&A. If the percentages have been fairly consistent
over the past three years, then this could be a good indication.
 BEST PRACTICES OF
FINANCIAL MODELLING
 1. Keep the Constituents Separate
• Separating the inputs, calculations, and outputs is a
mandatory requirement

• Being able to identify inputs of a spreadsheet is crucial for

understanding the effect on the outputs, such as what the
outputs are based on.
WAYS TO SEPARATE INPUTS, CALCULATIONS & OUTPUTS

Use different colors. For example, use a yellow cell fill for inputs
and a gray cell fill for calculations and outputs. This is usually
suitable for very small models.

Use different areas of a single worksheet. For example, label an

area “Inputs” at the top of the worksheet and “Calculations” below
the inputs with the outputs below the calculations. This is usually
suitable for relatively simple models.

Use different worksheets for the inputs, calculations, and outputs

for medium and large models.
2. Minimum Implicit Assumptions
Sometimes assumptions will get built into the structure of the
model. For example, an assumption that the inflation rate of 4%
will start from Year 2 could be implicit in having the repayment
calculations start from the second time period in the model.

Even so, a better way would be to model this assumption with an

explicit input so that the user can quickly see how much the
inflation rate is in each year and can vary the rate if needed. This
would make the formulas more complex, but the model would be
more adaptable.
3. Avoid Using Constants Inside Formulas
● It is best to steer clear from using constants inside formulas
as they are a cause of major errors in models. You should use
constants inside formulas only for very obvious things that
never change, such as there being 12 months in a year.

● Any less straightforward constants usually warrant being

separated out just to make clear that they exist.
4. Let Inputs Drive Your Models, Not Calculations

● The assumptions that you develop for the model should

be guided by the final users’ thinking.
● If there are many different types of assumptions, put
them on the assumption’s sheet in separate small
tables with their own headings. Create groups by using
tables and labels for sections and subsections.
● A good approach is to apply indentation to make the
logical hierarchy obvious or use grouping to separate
headings.
5. Have Only One Input for Each Assumption

● Duplicating input assumptions will require the user to

change a given input in several places. Critically, if the
user is unaware of all the changes that are required or
forgets to make the changes, then the model will have
some flawed inputs, dramatically increasing the risk of
error.
6. Specify Measurement Units for All Input
Assumptions
● Model users should be able to understand the measurement
unit for every input assumption. Modelers should not assume
that everyone who sees the model will be equally informed.
● To avoid confusion when a large number of units are involved,
designate a separate column for measurement units. This
step also helps to avoid errors with the conversion of units.
● Typically, the measurements can range from anything to
currencies, dates, volumes, or resources.
7. The Logical Flow of Calculations and Circularity

● Formulas should take their parameters from rows above and

columns to the left, as this makes the organization of the
model more logical.
● Always consider, if it is possible, to use a different approach
or to change the input assumptions to avoid circularities, as
they usually make models more difficult to use and update.
● Sometimes, it is preferable to create a macro that would
copy and paste values to break the circularity rather than to
allow circular references in Excel.
8. Use Consistent Formulas Across Rows

● When working with models, it is extremely important to use

consistent formulas across the rows. When this principle is
not followed, problems and errors will likely occur.
● For every calculation row, input a single formula and copy it
across all the columns. Avoid using a different formula
somewhere in the middle of a row.
● In case you are bound to do so, ensure you insert a comment
in this regard to warn the end users.
9. Break Up Long Formulas into Simple Pieces
● It is best not to reduce the number of cells used by trying to
condense too many calculations into a single formula.
Writing long and complex formulas that nobody can
understand is generally the mark of bad modeling practice.
● Try to instead split a complex calculation into smaller pieces
and utilize as many rows as required. Usually the more
separate calculations you use, the easier it will be to follow
and understand the model.
● With this format, more logical elements will be labeled and
the resulting formulas will be much simpler.
10. Include Automatic Error Checks
● A good quality financial model will always have various error
traps and checks built into its logic. These help to ensure
internal consistency of inputs, calculations, and outputs.
Most often such checks represent simple formulas that
return zero in case of success and a nonzero value in case of
error.
● For example, to check that a balance sheet balances across
all time periods, you should add a row that would subtract
total assets from total liabilities. It is a good idea to use
different formatting for that row, such as italics or a red font.
 Remember:
A good model is a functional and flexible
one, and is one that is designed to easily
be adjusted, to grow, and to evolve as
we gain more knowledge and insight
into the inner workings of business,
therefore slowly honing on a perfect
valuation.