Decision Trees

 Quantitative decision-making model

 Applied to problems for which the probability of different
consequences is known and the financial outcome of the
decision can be estimated

 It maps out different options and allows the best to be

chosen
Decision trees are a form of diagrammatic analysis
used to help businesses with making decisions
where there are a number of different options
from which to select.

Decision trees are particularly useful in situations

where chance (or probability) plays an important
role in likely outcomes.
 A decision tree is a Quantitative tool used to help
managers make decisions.

 A decision tree uses estimates and probabilities

to calculate likely outcomes.
 Decision trees are particularly helpful in
situations supporting complex business
decisions or problems, involving more than one
decision.
 A decision is constructed from left to right with
events laid out in the sequence which they
occur, but the calculation of financial results is
always from right to left
Squares represent decision points
which are under the control of the
business.
The lines that come out of each
square show all the available
options that can be selected.
 decision tree starts with a decisions node
(square).

 Decision nodes represent points where a

business has to make a choice between
alternatives.
 From this decision node we draw out lines
towards the right for each possible option or
choice, and write the description of that option
along the line.

 At the end of each line, we consider the likely

outcome of a decision.

 If the result of taking that decision is

uncertain, we draw a circle – a chance or
probability node.
AN EXAMPLE OF A BASIC
DECISION TREE
Calculations to NODE 1:
Understand the Decision CALCULATION OF EV
Tree
Step 1: Outcome x Probability

$20 m x 0.8 = $16 million

($2 m) x 0.2 = ($0.4 million)

Step 2: Weighted Average

The EV (or weighted average) is

shown by adding the two results
together

EV = $16 m + ($.04 million) =

$15.6 million
NODE 2: CALCULATION OF EV

Step 1: Outcome x Probability Step 2: Weighted Averaged

$30 million x 0.7 = $21 million $21 million + 1.8 million = $22.8 million
$6 million x 0.3 = $1.8 million
Product A
• To achieve an EV of $15.6 million, the business has
had to spend $10 million. Therefore the, cost must be
deducted from the EV to find the final profit. (net EV)
• $15.6 million - $10 million = $5.6 million (net EV)

Product B
• To achieve an EV of $22.8 million, the business had to
spend $15 million. Therefore the cost must be
deducted from the EV to find the final profit.
• $22.8 million - $15 million = $7.8 million (net EV)
 Circles represent (probability or chance)
nodes.

 These show various circumstances that

have uncertain outcomes.
 The lines that come out of each circle
denote possible outcomes of that
uncontrollable circumstance.
Chance nodes represent points at which
chance, or probability, plays a dominant role
and reflect alternatives over which the
business has (effectively) little control.

For example, if one of the options a business

has is to launch a new product, the possible
outcomes of the launch are excellent sales,
good sales, and poor sales.
The chance or probability node allows us to
represent likelihood of these alternative
outcomes.
• It is difficult to establish accurate probabilities
of an event, especially if that event is beyond
the control of the business itself.
• However, estimates of probability can be
achieved through market research or
experience.
• Nevertheless, the fact that probabilities may be
little more than educated guesses is clearly a
weakness of the decision tree process.
Each probability is based on judgement and
research evidence.

That is, through gathering of evidence, it

may be estimated that the probability of
achieving ‘good sales’ is 40% (represented
as 0.4).

The probabilities from each chance node

are added together to total 1 to take into
account the full range of possibilities that
could occur.
 One decision can lead to the need to
make another decision

(i.e. decision trees can have more than

one decision node throughout)

For example, the first decision might be to launch a

new product.
The second decision might be to go for a national or
regional launch. For each decision that needs to
be made, another square is drawn.
 A calculation is required of the average
outcome given the probabilities.
 This is a weighted average, because the
outcome is multiplied by the probability of
that result happening.
 In decision trees, these are referred to as
expected values (EVs) and are shown at
every probability node.
 Although doing nothing may not have an
immediate direct cost, it certainly may
have a negative outcome.

 A failing business left to carry on failing

may end up with bankruptcy of the
owner and liquidation of the business,
unless some miracle happens.
 Probability is

 The percentage chance or possibility that an

event will occur.

 Ranges between 1 (100%) and 0.

 If all the outcomes of an event are considered,

the total probability must add up to 1
The financial value of an outcome
calculated by multiplying the
estimated financial effect by its
probability
 The value to be gained from taking a decision.
 Net gain is calculated by adding together the
expected value of each outcome and
deducting the costs associated with the
decision
 High sales: (0.6 x £1,000,000) = £600,000
 Low sales: (0.4 x £750,000) = £300,000

 Total expected value = £900,000

 Net gain: £900,000 - £500,000 = £400,000

 High sales: (0.8 x £800,000) = £640,000
 Low sales: (0.2 x £500,000) = £100,000

 Total expected value = £740,000

 Net gain: £740,000 - £300,000 = £240,000

 Both options indicate a positive net gain,
suggesting that either would be better than
doing nothing.

However, launching the loyalty
card has a higher net gain &
looks the best option of the
two considered
 Economic growth rises Expected outcome
£300,000
0.7

Expand by opening new outlet

Economic growth declines Expected outcome
-£500,000
0.3

Maintain current status

£0

A simple decision tree

The circle denotes the point where different outcomes could occur. The
The probability
estimates of the probability and expected outcome of allow
each aoutcome
calculation of
A square denotes the point where a decision is made.
the likely return. In thisisexample
drawn on the tree
it is:
Thereisiscontemplating
A business also the optionopening
to do nothing and maintain
a new outlet. the current
The uncertainty is the
Economic growth
status quo! rises:
This would0.7 x £300,000
have an = £210,000
outcome of
state of the economy – Economic growth is estimated to £0.
yield profits of
Economic growth declines: £300,000.
0.3 x £500,000 = -£150,000
If economy
The fails would
calculation to grow as expected,
suggest thetopotential
it is wise go ahead loss is the
with estimated at
decision
£500,000.
(an ‘expected value’ of +£60,000)
 The Process

Economic growth rises Expected

outcome
0.5 £300,000

Expand by opening new

outlet Economic growth declines Expected
outcome
0.5 -£500,000
Maintain current
status
£0

Look what happens however if the probabilities change. If the firm is

unsure of the potential for growth, it might estimate it at 50:50. In this
case the outcomes will be:
Economic growth rises: 0.5 x £300,000 = £150,000
Economic growth declines: 0.5 x -£500,000 = -£250,000
In this instance, the expected value is -£100,000
The decision looks less favourable!
 OUTCOME ACTIVITY
$30 m
0.7 Using the
information in
this diagram,
complete
($20 million) appropriate
calculations to
$3m determine
0.3 which product
should be
$40 m launched by the
0.6 company.
($30 million) PRODUCT A
Or
PRODUCT B??

$1 m
0.4
A business is choosing whether to invest in updating a product to
extend its product life cycle, or whether to let the product go into
decline. The information and research that the business has in
relation to this decision is in the table below:
if the business were to decide to invest in the
product, there would be a probability of 0.4 or
40% that this decision would lead to a return of
£120 000.

Alternatively if the business decided to allow the

product to decline, there is probability of 0.5
or 50% that this would lead to a return of £40
000.

The cost of the potential investment on the

decision tree (£30 000
Evaluative considerations

The predicted outcomes and estimates

need to be based on valid data and research
if they are to have any real quantitative
value.
As with all decision-making methods, decision
tree analysis should be used in conjunction with
other types of decision-making, decision trees are
just one important part of a business decision-making
tool kit.
Qualitative factors will also have an influence on
the decision made.

In the example above, qualitative factors such as

the effects of decisions on stakeholders e.g.
workforce, management, suppliers and customers
as well as training costs, recruitment, capacity
management, marketing impact and so on, all
need to be considered before a final decision is
made
 Once the value
Economic of each
growth £300,000
rises option has
been calculated
0.7 it is added to the
Expand by opening new decision tree
outlet
-£20,000 Economic growth declines -£500,000
Net = +£40,000 0.3
Cross through any
// rejected decisions
Do nothing £0
To calculate the expected value: Remember to subtract the initial
New outlet: 0.7 x £300,000 = £210,000
cost of the project to get the final
: 0.3 x -£500,000 = -£150,000 expected value (or ‘net benefit’)
= £ 60,000

= £ 60,000 - £20,000 = £40,000

 High demand
£16m
0.7
Product A Low demand
£6m
-£7m 0.3

High demand
Product B £12m
0.6
-£2m
Do nothing
Low demand
£4m
0.4
£0

 Most decision trees will have more than one option

 The tree above shows the choice a company will face between choosing to
produce product A or B
 To choose between them work out the expected value for both and choose the
most beneficial Which option should be chosen?
 Product A:
 0.7 x £16m = £11.2m
 0.3 x £6m = £1.8m
 Expected value = £11.2m + £1.8m = £13m - £7m = £6m

 Product B:
 0.6 x £12m = £7.2m
 0.4 x £4m = £1.6m
 Expected value = £7.2m + £1.6m = £8.8m - £2m = £6.8m

 B should be chosen, but the difference is narrow so qualitative factors

will be important
 New competitor £1.5m
AQA P.492 PE 2
0.7

launch No competitor £2.5m

-£1.2m
EV = £0.6m 0.3

//
Do nothing New competitor
-(£0.4m)
EV = £0.2m 0.4

No competitor
Expected value (Do nothing):
0.4 x –£0.4m = -£0.16m
0.6 x £0.6m = £0.36m
0.16 + 0.36 = £0.2m - £0 = £0.2m
£0.6m
0.6
Expected value (Launch):
0.7 x 1.5 = £1.05m
0.3 x 2.5 = £0.75m
1.05 - 0.75 = £1.8m - £1.2m = £0.6m
Ex4 p.493

On financial grounds alone, the move to cheaper premises is more desirable

 Encourage a careful consideration of all alternatives
 Set out problems clearly and give a logical approach
 A quantitative (numerical) consideration of chance is adopted
 Risk is taken into account
 Discourages hunch decisions
 Choices are set out in a logical way
 Potential options & choices are considered at
the same time
 Use of probabilities enables the “risk” of the
options to be addressed
 Likely costs are considered as well as potential
benefits
 Easy to understand & tangible results
 Reliant on the accuracy of the data used
 Qualitative factors not considered – human resources, motivation,
reaction, relations with suppliers and other stakeholders
 Difficult to get accurate probabilities as estimated. What market
research has been done?
 Requires qualitative input to give complete picture

Qualitative issues must not be ignored just because the

expected value looks convincing

Competitor actions, future recession?

 Probabilities are just estimates – always prone
to error
 Uses quantitative data only – ignores
qualitative aspects of decisions
 Assignment of probabilities and expected
values prone to bias
 Decision-making technique doesn’t
necessarily reduce the amount of risk
 Complete A-Z questions worksheet (25 minutes)
A-Z Activity decision tree
Max contribution £500,000

Electric Medium £200,000

Disappointment
£80,000

Gas
£150,000

£80,000
1. What is a decision tree?

2. Explain 2 advantages of using it

3. Explain 2 disadvantages of using it

4. Why might it be less useful concerning new

situations?

5. Explain 3 qualitative factors that might be important.

6. If the probability of success is o.6, what is the

probability of failure?