Decision Trees: Recap/Do Now
Decision Trees: Recap/Do Now
Decision trees
Recap/Do Now
• Construct simple decision trees
• Interpret a decision tree
• Explain limitations of using DT
Definition
A decision tree represents
the likely financial outcomes
for a business from different
courses of action
Drawing a decision tree..(options and costs)
Develop a new chicken burger
£5000
Point £10,000
Do nothing
Add possible outcomes and probabilities (add up
to 1)
High sales
Develop a new burger 0.8
£5000 1 Low sales
0.2
High sales
Decision Advertise current burger 0.6
£10,000
2 Low sales
0.4
Do nothing
Add financial outcomes
High sales
£20,000
Develop a new chicken burger 0.8
£5000 1 Low sales
£5,000
0.2
High sales
Decision £30,000
Advertise current burgers 0.6
£10,000
2 Low sales
£10,000
0.4
Do nothing
Calculations (EV and net gain)
High sales
£20,000
Develop a new chicken burger 0.8 £20,000 x 0.8 = 16,000
£5000 1 Low sales
£5,000
Net gain = EMV = 0.2
12,000 17,000 £5,000 x 0.2 = 1,000
High sales
Decision £30,000
Advertise current burgers 0.6 £30,000 x 0.6 = 18,000
£10,000
2 Low sales
£10,000
Net gain = 12,000 EMV = 0.4 £10,000 x 0.4 = 4,000
Do nothing 22,000
Another one…
Exercise – create the tree, work out EV and net
gain for both
Scenario
BIG BURGERS operates 50 burger shops in and around London. It is looking at two options
to increase revenues across the chain. The estimated impact of the two options on sales
are shown below
Is useful... However...
5 mins
to plan
your an
swer an
d1 0 mins
to answ
er i t
Short term...
Long term...
It depends...
1 1-2 e.g. Decision trees help businesses to attach probabilities to the outcomes of decisions
e.g. Expected outcomes/values give a financial value to the results of decisions which
account for different probabilities
2 3-4 e.g. Toyota could estimate the financial cost of recalls in different markets
e.g. Toyota is an established business and will have lots of previous sales data to help
construct probabilities and expected outcomes
3 5-6 e.g. Decision tree analysis will help Toyota to quantify their recall decision as this can
be more easily communicated to key stakeholders such as managers e.g. If the full
financial implications of the recall decision can be estimated accurately this means
decision trees may be a rational basis for making decisions
e.g. Decision trees take time to construct accurately and customers need immediate
feedback on the status of their vehicle
4 7-10 e.g. Decision trees are based on predicted data of the potential impact of a decision
such as a recall
e.g. Decision tree does not take into account unforeseen costs and circumstances such
as a delay in shipping.
e.g. Using decision trees might fit with the cautious culture exhibited by the business
but this is likely to be overridden by the need to act quickly, which would suggest the
decision trees are not that useful in this instance
Kahoots
Plenary – Pick a box
1 2 Ask someone a question 3 Lucky box
based on today’s lesson You get to choose someone
Define decision tree else to pick a box
Very lucky box How to calculate the How to calculate the net gain
7You get to choose someone
else to pick a box and you
8 expected value? 9
can’t be picked again
• Come up with 5 top tips to remember today’s learning