Decision Tree (2nd Day)

Choice of the Best Alternative

Syndicate 2
Dina Rizkia Rachmah (29120431)
Muhammad Ikhlas Dharma (29120398)
Desman Hansen Sagala (29120481)
Table of Contents

01 02
Exercise 3 Conclusion
Case on Budi as a Conclusion of the case
marketing manager

2
 01
Exercise 3
Case on Budi as a marketing
manager

3
Budi was a marketing manager. He was assigned to assess the
prospect of a new product to be manufactured. The approximately
production cost was USD 90,000. If it had been sold out then it would
gain revenue USD 500,000. In other side, if he didn’t decide to
produce then the company would loss nothing. Based on his
experience, he knew that 0.2 was the probability that the market
would accept the product. Even though, he wasn’t sure whether it
represented the current data. He had another option to run market
research and collect the last information, but it took investments as
many as USD 10,000. Again, based on experience, he had known
that if the customer preferred the product then the survey would
exhibit that the result was favorable with probability 0.6. However,
the probability would reduce to 0.1, if the customer rejected.
—Now, help Budi to decide the best strategy! …. and … don’t forget to :

1. Draw the tree diagram.

2. Show us how did you get the probability.
4
The Overall Tree Diagram

Favorable

Do survey

Unfavorable

Accept
Produce

Reject

No survey
No produce
5
 Budi as a Marketing Manager

The approximately production cost was USD 90,000. If it had been sold out then it
would gain revenue USD 500,000. Approximate profit of:

Profit if the market would accept the product = Revenue if Sold out – Production cost
= $500,000 - $90,000
= $410,000

In other side, if he didn’t decide to produce then the company would loss nothing.

6
● Decision that must be taken:
Should Budi Produce the Product or Do not produce?

● Alternatives:
-Produce the new Product
-Do not produce the new product

● Possibility event that could happen (State of Nature)

-Market would Accept the Product
-Market would Reject the Product

● Payoff Table (Information from the case)

State of Nature

Alternative Accept Rejected EV

Produce $ 410.000 -$90.000 $10.000

No Produce $0 $0 0

Prior 0.2 0.8

Probability 7
Decision Criterion
1. MAXIMAX (Maximum of Maximum ) = $410,000 , Produce!
2. MAXIMIN (Maximum of Minimum) = $ 0 No Produce!
3. Likelihood (High Probability) = 0.8 No Produce!
4. Bayes > Expected Value (EV)

EV(Produce) = Sigma p(Event) * Payoff

= (0.2)*(410,000) + (0.8)*(-90,000)
= $ 10,000

EV(No Produce) = Sigma p(Event) * Payoff

= (0.2)*(0) + (0.8)*(0)
=0
Expected Payoff :
= Max{EV(Produce),EV(No Produce)}
= Max{10,000 , 0} = 10,000

Action : Produce !
8
The No Survey Tree Diagram
Accept $410,000 (0.2)

Produce

$10,000

Reject -$90,000 (0.8)

No survey

Accept $0 (0.2)

$0

No produce

Reject $0 (0.8)

9
Budi’s Case Continued
Budi wasn’t sure whether it represented the current data. He had another option to run
market research(Survey) and collect the last information.

Survey by a marketing manager will provide more accurate information about

P(Accept);

How if Budi has to decide 2 alternatives:

1. Do survey before Produce/ No produce
2. Produce/ No Produce without Survey

Events:
Do Survey
● Fav :Favorable Market Research
● Unfav : Unfavorable Market Research

Produce or No Produce
● Accepted
● Rejected
10
 Budi’s Experience
1. P(state) → Prior ; Which is P(Accepted)=0,2 & P(Rejected) = 0,8
2. P(finding|state) → being known based on Budi’s Experience; Which is

P(Fav|Accepted) = 0,6 ; so →P(Unfav|Accepted) = 1-(0,6) = 0,4

P(Fav|Rejected) = 0,1 ; so →P(Unfav|Rejected) = 1-(0,1) = 0,9

3. Calculate value of P (Fav), P(Unfav), P(Fav and Accepted), P(Fav and Rejected),

P(Unfav and Accepted) also P(Unfav and Rejected). → [Posterior Probability Formula]

P(Fav) = P(Fav and Accepted) + P(Fav and Rejected) → [Marginal Probabilities]

P(A|B) = P(A and B) / P(B) → [Conditional Probability]

P(Fav|Accepted) = P(Fav and Accepted) / P(Accepted) P(Fav|Rejected) = P(Fav and Rejected) / P(Rejected)

0,6 = P(Fav and Accepted) / 0,2 0,1 = P(Fav and Rejected) / 0,8

P(Fav and Accepted) = (0,6)*(0,2) = 0,12 P(Fav and Rejected) = (0,1)*(0,8) = 0,08

P(Fav) = P(Fav and Accepted) + P(Fav and Rejected) = (0,12) + (0,08) = 0,2 11
Step 3- Sequel or can just make a table for probability calculation.

P(Unfav|Accepted) = P(Unfav and Accepted) / P(Accepted) P(Unfav|Rejected) = P(Unfav and Rejected) / P(Rejected)

0,4 = P(Unfav and Accepted) / 0,2 0,9 = P(Unfav and Rejected) / 0,8

P(Unfav and Accepted) = (0,4)*(0,2) = 0,08 P(Unfav and Rejected) = (0,9)*(0,8) = 0,72

P(Unfav) = P(Unfav and Accepted) + P(Unfav and Rejected) = (0,08) + (0,72) = 0,8

4. Make a Contingency Table for probability calculation:

Accepted Rejected Probability

Favorable 0,12 0,08 0,2

Unfavorable (0,20 - 0,12) = 0,08 (0,80 - 0,08) = 0,72 1 - (0,2) = 0,8

Probability 0,20 0,80 1

12
 5. Calculate the the other Posterior Probability

a. P(Accepted|Fav) & P(Rejected|Fav)

b. P(Accepted|Unfav) & P(Rejected|Unfav)

Law of Probability → P(Fav and Accepted) = P(Accepted and Fav)

P(Accepted|Fav) = P(Accepted and Fav) / P(Fav) P(Accepted|Unfav) = P(Accepted and Unfav) / P(Unfav)
P(Accepted|Fav) = 0,12 / 0,2 P(Accepted|Unfav) = 0,08 / 0,8
P(Accepted|Fav) = 0,6 P(Accepted|Unfav) = 0,1

P(Rejected|Fav) = P(Rejected and Fav) / P(Fav) P(Rejected|Unfav) = P(Rejected and Unfav) / P(Unfav)
P(Rejected|Fav) = 0,08 / 0,2 P(Rejected|Unfav) = 0,72 / 0,8
P(Rejected|Fav) = 0,4 P(Rejected|Unfav) = 0,9
or or
P(Rejected|Fav) = 1 - 0,6 = 0,4 P(Rejected|Unfav) = 1 - 0,1 = 0,9

13
Leveled Decision Analysis

EV1.1.1.1

EV1.1.1 *note:
multiplication result
EV1.1.1.2
choose the optimal
EV1.1
EV1.1.2

EV1 EV1.2.1.1

EV1.2.1 EV1.2.1.2

EV1.2
EV1.2.2

EV2.1.1

EV2.1 EV2.1.2
EV2

EV2.2

14
 Expected Value Calculation

EV1.1.1= EV1.1.1.1 + EV1.1.1.2= (0,6*400.000)+(0,4*(-100.000))= (240.000 +(-40.000))= 200.000

EV1.1.2= Research cost -> -10.000
EV1.2.1= EV1.2.1.1 + EV1.2.1.2= (0,1*400.000)+(0,9*(-100.000))= (40.000+(-90.000))= -50.000
EV1.2.2= Research cost -> -10.000
EV2.1= EV2.1.1 + EV2.1.2= (0,2*410.000)+(0,8*+(-90.000))= 10.000

After multiplication process we got the optimal result, those are: EV1.1.1, EV1.2.2, and EV2.1.
Because of EV1 oval symbol, we should use multiplication process between EV1.1.1 and EV1.2.2
EV1= EV1.1.1 + EV1.2.2= (0,2*200.000)+(0,8*(-10.000))= 32.000

The last, we should choose the optimal strategy. According to tree diagram there are two
strategy which are EV1 and EV2. Then, The best strategy is EV1 because gain revenue 32.000.

15
 02
Conclusion
Conclusion of the case

16
 What Budi
should do:
Conduct a survey. If the result of the survey
obtained is unfavorable → do not produce.
Whereas, if the result of the survey obtained
if favorable → do produce.

17
 Thanks!
Do you have any questions?

CREDITS: Diese Präsentationsvorlage wurde von

Slidesgo erstellt, inklusive Icons von Flaticon und
Infografiken & Bilder von Freepik

18