Decision Theory and Decision Trees in Marketing

Marketing managers must constantly make high-stakes decisions under conditions of risk and incomplete information. Should a company launch a new product nationwide, test market it first, or abandon the project? Decision Theory and Decision Trees provide structured mathematical and logical frameworks to evaluate alternative choices and their financial payoffs.

1. Decision-Making Environments

In decision theory, the operational environment is classified based on the degree of certainty:

   ┌────────────────────────────────────────────────────────┐
   │ DECISION-MAKING ENVIRONMENTS:                          │
   │ 1. Decision Making under CERTAINTY (Known outcomes)    │
   │ 2. Decision Making under RISK (Known probabilities)    │
   │ 3. Decision Making under UNCERTAINTY (Unknown probs)   │
   └────────────────────────────────────────────────────────┘

Under Certainty: The manager knows with 100% certainty what state of nature will occur (rare in marketing).
Under Risk: The manager does not know the exact outcome, but can assign objective statistical probabilities to various market events (e.g., 60% chance of high economic growth, 40% chance of recession).
Under Uncertainty: Several market outcomes are possible, but the manager has zero historical data or basis to assign even subjective probabilities.

2. Decision Criteria Under Uncertainty

When probabilities cannot be assigned, managers apply non-probabilistic decision criteria based on risk personality:

Decision Criterion	Philosophy	Managerial Mindset
Maximax (Optimistic / Hurwicz)	Select the decision alternative that maximizes the maximum possible payoff.	Aggressive risk-taker: “Go for the absolute highest jackpot!”
Maximin (Pessimistic / Wald)	Identify the worst possible payoff for each alternative and pick the “best of the worst”.	Ultra-conservative: “Protect against catastrophic losses at all costs.”
Laplace (Equal Probability)	Assume all market states of nature are equally likely and select the highest average payoff.	Rationalist: “If we have no data, treat all events as having equal odds.”
Minimax Regret (Savage)	Calculate the “opportunity loss / regret” for each choice and minimize the maximum regret.	Regret-avoider: “Ensure I never look back and regret a missed opportunity.”

3. Decision Making Under Risk: Expected Monetary Value (EMV)

When probabilities can be estimated, the standard tool is Expected Monetary Value (EMV).

EMV Calculation Formula:

The EMV of an alternative is the weighted sum of all possible financial payoffs multiplied by their respective probabilities:

   EMV = (Payoff 1 x Prob 1) + (Payoff 2 x Prob 2) + ... + (Payoff n x Prob n)

The rational decision rule is to select the alternative with the highest EMV.

Expected Value of Perfect Information (EVPI):

How much money should a company be willing to pay a marketing research agency for guaranteed perfect market predictions?
EVPI = Expected Value with Perfect Information (EVwPI) - Maximum EMV without Information.
EVPI sets the theoretical upper spending limit for any market research study.

4. Decision Trees in Marketing

A Decision Tree is a visual diagram that maps out sequential decision alternatives, uncertain market events, associated probabilities, and final financial payoffs.

Core Symbols of a Decision Tree:

   ┌─────────┐
   │         │  DECISION NODE (Square):
   │    □    │  Represents a point where the manager must choose an action.
   └─────────┘

      ╭───╮
     │  ○  │     CHANCE / EVENT NODE (Circle):
      ╰───╯      Represents an uncertain market event beyond managerial control.

   ───────────►  BRANCH: Represents an alternative path or event.

   [ Rs. Payoff ] TERMINAL OUTCOME: The final net profit or loss at the branch tip.

Example: New Product Launch Decision Tree

                                      ┌── High Demand (p=0.6) ──► Profit: Rs. 10 Lakhs
                  ┌── Launch Directly─┤
                  │                   └── Low Demand  (p=0.4) ──► Loss: -Rs. 4 Lakhs
                  │
   [Decision Node]┤
   (Choose Path)  │                   ┌── Positive Test (p=0.7) ─► Launch ─► Profit: Rs. 7 Lakhs
                  ├── Test Market ────┤
                  │   (Cost: Rs. 1 L) └── Negative Test (p=0.3) ─► Abort  ─► Loss: -Rs. 1 Lakh
                  │
                  └── Do Not Launch ────────────────────────────► Payoff: Rs. 0

Solving a Decision Tree (Backward Roll-Back Process):

Draw the tree from left to right: Lay out decision choices, chance events, probabilities, and terminal payoffs.
Solve the tree from right to left (Roll-back):
- At each Chance Node (Circle): Calculate the EMV by multiplying payoffs by their branch probabilities.
- At each Decision Node (Square): Compare the calculated EMVs of all emerging branches and select the highest EMV, pruning away inferior choices.

Summary Checklist for Students

Defined Decision Making under Certainty, Risk, and Uncertainty.
Mastered the 4 criteria under uncertainty: Maximax, Maximin, Laplace, and Minimax Regret.
Understood how to calculate Expected Monetary Value (EMV) and EVPI.
Identified Decision Tree building blocks: Square (decision), Circle (chance event), and Terminal Payoff.
Learned the Backward Roll-Back method to solve marketing decision trees.